Train 1D CNN

1D Convolution을 기본 구성 요소로 하는 EEG classifier를 학습해보는 노트북.

• Eye-open cropping
• Three-way SoftMax classifier of normal, non-vascular MCI, and non-vascular dementia

---

환경 구성

# for auto-reloading external modules
# see http://stackoverflow.com/questions/1907993/autoreload-of-modules-in-ipython
%load_ext autoreload
%autoreload 2

# Load some packages
import os
import glob
import json

import matplotlib.pyplot as plt
import pprint
from IPython.display import clear_output
from tqdm.auto import tqdm

import numpy as np
import random
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torch.utils.data import DataLoader
from torchvision import transforms
from typing import Type, Any, Callable, Union, List, Optional

# custom package
from utils.eeg_dataset import *

# Other settings
%matplotlib inline
%config InlineBackend.figure_format = 'retina' # cleaner text

plt.style.use('default') 
# ['Solarize_Light2', '_classic_test_patch', 'bmh', 'classic', 'dark_background', 'fast', 
#  'fivethirtyeight', 'ggplot', 'grayscale', 'seaborn', 'seaborn-bright', 'seaborn-colorblind', 
#  'seaborn-dark', 'seaborn-dark-palette', 'seaborn-darkgrid', 'seaborn-deep', 'seaborn-muted', 
#  'seaborn-notebook', 'seaborn-paper', 'seaborn-pastel', 'seaborn-poster', 'seaborn-talk', 
#  'seaborn-ticks', 'seaborn-white', 'seaborn-whitegrid', 'tableau-colorblind10']

plt.rcParams['image.interpolation'] = 'nearest'
plt.rcParams["font.family"] = 'NanumGothic' # for Hangul in Windows

print('PyTorch version:', torch.__version__)
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

if torch.cuda.is_available(): print('cuda is available.')
else: print('cuda is unavailable.')

PyTorch version: 1.9.0
cuda is available.

# Data file path
root_path = r'dataset/'

meta_path = os.path.join(root_path, 'metadata_debug.json')
with open(meta_path, 'r') as json_file:
    metadata = json.load(json_file)

pprint.pprint(metadata[0])

{'age': 78,
 'birth': '1940-06-02',
 'dx1': 'mci_rf',
 'edfname': '00001809_261018',
 'events': [[0, 'Start Recording'],
            [0, 'New Montage - Montage 002'],
            [36396, 'Eyes Open'],
            [72518, 'Eyes Closed'],
            [73862, 'Eyes Open'],
            [75248, 'Eyes Closed'],
            [76728, 'swallowing'],
            [77978, 'Eyes Open'],
            [79406, 'Eyes Closed'],
            [79996, 'Photic On - 3.0 Hz'],
            [80288, 'Eyes Open'],
            [81296, 'Eyes Closed'],
            [82054, 'Photic Off'],
            [84070, 'Photic On - 6.0 Hz'],
            [84488, 'Eyes Open'],
            [85538, 'Eyes Closed'],
            [86086, 'Photic Off'],
            [88144, 'Photic On - 9.0 Hz'],
            [90160, 'Photic Off'],
            [91458, 'Eyes Open'],
            [92218, 'Photic On - 12.0 Hz'],
            [92762, 'Eyes Closed'],
            [94198, 'Photic Off'],
            [94742, 'Eyes Open'],
            [95708, 'Eyes Closed'],
            [96256, 'Photic On - 15.0 Hz'],
            [98272, 'Photic Off'],
            [100330, 'Photic On - 18.0 Hz'],
            [102346, 'Photic Off'],
            [102596, 'Eyes Open'],
            [103856, 'Eyes Closed'],
            [104361, 'Photic On - 21.0 Hz'],
            [106420, 'Photic Off'],
            [106880, 'Eyes Open'],
            [107804, 'Eyes Closed'],
            [108435, 'Photic On - 24.0 Hz'],
            [110452, 'Photic Off'],
            [111080, 'Eyes Open'],
            [112004, 'Eyes Closed'],
            [112509, 'Photic On - 27.0 Hz'],
            [114528, 'Photic Off'],
            [114864, 'Eyes Open'],
            [116124, 'Eyes Closed'],
            [116544, 'Photic On - 30.0 Hz'],
            [118602, 'Photic Off'],
            [126672, 'artifact'],
            [134030, 'Move'],
            [135584, 'Eyes Open'],
            [136668, 'Eyes Closed'],
            [139818, 'Eyes Open'],
            [141414, 'Eyes Closed'],
            [145000, 'Paused']],
 'label': ['mci', 'mci_amnestic', 'mci_amnestic_rf'],
 'record': '2018-10-26T15:46:26',
 'serial': '00001'}

---

Data Filtering by Event

print('Before:', len(metadata))

# No eyes open
out_list = []
for (i, m) in enumerate(metadata):
    num = 0
    for e in m['events']:
        if e[1].lower() == 'eyes open':
            num += 1
    if num == 0:
        out_list.append(i)
        
# insufficient time before the eyes open event
before = 2000
for (i, m) in enumerate(metadata):
    num = 0
    for e in m['events']:
        if e[1].lower() == 'eyes open' and e[0] > before:
            num += 1
    if num == 0:
        out_list.append(i)
        
metadata = [m for i, m in enumerate(metadata) if i not in out_list]

print('After:', len(metadata))

Before: 1387
After: 1383

---

Data Filtering by Diagnosis

Non-Vascular Dementia, Non-Vascular MCI, Normal

diagnosis_filter = [
    # Normal
    {'type': 'Normal',
     'include': ['normal'], 
     'exclude': []},
    # Non-vascular MCI
    {'type': 'Non-vascular MCI',
     'include': ['mci'], 
     'exclude': ['mci_vascular']},
    # Non-vascular dementia
    {'type': 'Non-vascular dementia',
     'include': ['dementia'], 
     'exclude': ['vd']},
]

def generate_class_label(label):
    for c, f in enumerate(diagnosis_filter):
        # inc = set(f['include']) & set(label) == set(f['include'])
        inc = len(set(f['include']) & set(label)) > 0        
        exc = len(set(f['exclude']) & set(label)) == 0
        if  inc and exc:
            return (c, f['type'])
    return (-1, 'The others')

class_label_to_type = [d_f['type'] for d_f in diagnosis_filter]
print('class_label_to_type:', class_label_to_type)

class_label_to_type: ['Normal', 'Non-vascular MCI', 'Non-vascular dementia']

splitted_metadata = [[] for i in diagnosis_filter]

for m in metadata:
    c, n = generate_class_label(m['label'])
    if c >= 0:
        m['class_type'] = n
        m['class_label'] = c
        splitted_metadata[c].append(m)
        
for i, split in enumerate(splitted_metadata):
    if len(split) == 0:
        print(f'(Warning) Split group {i} has no data.')
    else:
        print(f'- There are {len(split):} data belonging to {split[0]["class_type"]}')

- There are 461 data belonging to Normal
- There are 347 data belonging to Non-vascular MCI
- There are 228 data belonging to Non-vascular dementia

---

Configure the Train, Validation, and Test Splits

Split the filtered dataset and shuffle them

# random seed
random.seed(0)

# Train : Val : Test = 8 : 1 : 1
ratio1 = 0.8
ratio2 = 0.1

metadata_train = []
metadata_val = []
metadata_test = []

for split in splitted_metadata:
    random.shuffle(split)
    
    n1 = round(len(split) * ratio1)
    n2 = n1 + round(len(split) * ratio2)

    metadata_train.extend(split[:n1])
    metadata_val.extend(split[n1:n2])
    metadata_test.extend(split[n2:])

random.shuffle(metadata_train)
random.shuffle(metadata_val)
random.shuffle(metadata_test)

print('Train data size\t\t:', len(metadata_train))
print('Validation data size\t:', len(metadata_val))
print('Test data size\t\t:', len(metadata_test))

print('\n', '--- Recheck ---', '\n')
train_class_nums = np.zeros((len(class_label_to_type)), dtype=np.int32)
for m in metadata_train:
    train_class_nums[m['class_label']] += 1

val_class_nums = np.zeros((len(class_label_to_type)), dtype=np.int32)
for m in metadata_val:
    val_class_nums[m['class_label']] += 1

test_class_nums = np.zeros((len(class_label_to_type)), dtype=np.int32)
for m in metadata_test:
    test_class_nums[m['class_label']] += 1

print('Train data label distribution\t:', train_class_nums, train_class_nums.sum())
print('Val data label distribution\t:', val_class_nums, val_class_nums.sum())
print('Test data label distribution\t:', test_class_nums, test_class_nums.sum())

# random seed
random.seed()

# print([m['serial']  for m in metadata_train[:15]])
# print([m['serial']  for m in metadata_val[:15]])
# print([m['serial']  for m in metadata_test[:15]])

Train data size		: 829
Validation data size	: 104
Test data size		: 103

 --- Recheck --- 

Train data label distribution	: [369 278 182] 829
Val data label distribution	: [46 35 23] 104
Test data label distribution	: [46 34 23] 103

Wrap the splitted data using PyTorch Dataset

ages = []
for m in metadata_train:
    ages.append(m['age'])

ages = np.array(ages)
age_mean = np.mean(ages)
age_std = np.std(ages)

print('Age mean and standard deviation:')
print(age_mean, age_std)

Age mean and standard deviation:
69.98431845597105 9.77458831701377

composed = transforms.Compose([EEGNormalizeAge(mean=age_mean, std=age_std),
                               EEGDropPhoticChannel(),
                               EEGEyeOpenCrop(crop_before=200*2, crop_after=200*2, jitter=20, mode='random'), # 4s in total
                               EEGToTensor()])

train_dataset = EEGDataset(root_path, metadata_train, composed)


composed = transforms.Compose([EEGNormalizeAge(mean=age_mean, std=age_std),
                               EEGDropPhoticChannel(),
                               EEGEyeOpenCrop(crop_before=200*2, crop_after=200*2, jitter=3, mode='random'), # 4s in total
                               EEGToTensor()])

val_dataset = EEGDataset(root_path, metadata_val, composed)
test_dataset = EEGDataset(root_path, metadata_test, composed)

print(train_dataset[0]['signal'].shape)
print(train_dataset[0])

print()
print('-' * 100)
print()

print(val_dataset[0]['signal'].shape)
print(val_dataset[0])

print()
print('-' * 100)
print()

print(test_dataset[0]['signal'].shape)
print(test_dataset[0])

torch.Size([20, 800])
{'signal': tensor([[-28., -27., -26.,  ...,  89., 104.,  96.],
        [ 26.,  28.,  27.,  ...,  50.,  49.,  47.],
        [  0.,   0.,   0.,  ...,  10.,   4.,   4.],
        ...,
        [-10.,  -8.,  -8.,  ...,  10.,   8.,   7.],
        [ -4.,  -7.,  -8.,  ...,  -6.,  -7.,  -6.],
        [ -8.,  45.,  -4.,  ...,  13.,  -2., -38.]]), 'age': tensor(-1.2261), 'class_label': tensor(0), 'metadata': {'serial': '01011', 'edfname': '01212635_140515', 'birth': '1956-06-01', 'record': '2015-04-14T11:11:58', 'age': 58, 'dx1': 'cb_normal', 'label': ['normal', 'cb_normal'], 'events': [[0, 'Start Recording'], [0, 'New Montage - Pz Montage'], [440, 'Eyes Open'], [5774, 'Eyes Closed'], [12788, 'Eyes Open'], [18878, 'Eyes Closed'], [23898, 'Eyes Open'], [29377, 'Eyes Closed'], [36392, 'Eyes Open'], [41936, 'Eyes Closed'], [48656, 'Eyes Open'], [54116, 'Eyes Closed'], [61046, 'Eyes Open'], [66254, 'Eyes Closed'], [72218, 'Eyes Open'], [77426, 'Eyes Closed'], [84944, 'Eyes Open'], [89858, 'Eyes Closed'], [96326, 'Eyes Open'], [101324, 'Eyes Closed'], [108926, 'Eyes Open'], [113714, 'Eyes Closed'], [120000, 'Paused']], 'class_type': 'Normal', 'class_label': 0}}

----------------------------------------------------------------------------------------------------

torch.Size([20, 800])
{'signal': tensor([[  7.,   8.,   8.,  ...,  19.,  17.,  20.],
        [  5.,   6.,   7.,  ...,   2.,   3.,   3.],
        [ -1.,  -2.,  -2.,  ...,   7.,   9.,   7.],
        ...,
        [  5.,   2.,  -1.,  ...,   3.,   1.,  -2.],
        [  4.,  -1.,  -7.,  ...,   1.,   0.,  -3.],
        [  3.,   1.,  11.,  ..., 122., 119., 115.]]), 'age': tensor(0.7177), 'class_label': tensor(1), 'metadata': {'serial': '00700', 'edfname': '00985401_011117', 'birth': '1940-09-09', 'record': '2017-11-01T14:20:48', 'age': 77, 'dx1': 'mci amnestic', 'label': ['mci', 'mci_amnestic'], 'events': [[0, 'Start Recording'], [0, 'New Montage - Montage 002'], [1705, 'Eyes Open'], [5402, 'Eyes Closed'], [13046, 'Eyes Open'], [17666, 'Eyes Closed'], [30308, 'Eyes Closed'], [36272, 'Eyes Open'], [41774, 'Eyes Closed'], [48958, 'Eyes Open'], [55510, 'Eyes Closed'], [61641, 'Eyes Open'], [66766, 'Eyes Closed'], [72730, 'Eyes Open'], [78988, 'Eyes Closed'], [87770, 'Eyes Open'], [90542, 'Eyes Closed'], [97096, 'Eyes Open'], [102178, 'Eyes Closed'], [110872, 'Eyes Open'], [113728, 'Eyes Closed'], [122052, 'Photic On - 3.0 Hz'], [122428, 'Eyes Open'], [123309, 'Eyes Closed'], [124068, 'Photic Off'], [126126, 'Photic On - 6.0 Hz'], [128142, 'Photic Off'], [130158, 'Photic On - 9.0 Hz'], [132216, 'Photic Off'], [132718, 'Eyes Open'], [133600, 'Eyes Closed'], [134232, 'Photic On - 12.0 Hz'], [136248, 'Photic Off'], [138306, 'Photic On - 15.0 Hz'], [140322, 'Photic Off'], [142380, 'Photic On - 18.0 Hz'], [142630, 'Eyes Open'], [143302, 'Eyes Closed'], [144396, 'Photic Off'], [146412, 'Photic On - 21.0 Hz'], [148428, 'Photic Off'], [150486, 'Photic On - 24.0 Hz'], [152502, 'Photic Off'], [152710, 'Eyes Open'], [153550, 'Eyes Closed'], [154560, 'Photic On - 27.0 Hz'], [156576, 'Photic Off'], [158592, 'Photic On - 30.0 Hz'], [160608, 'Photic Off'], [160942, 'Eyes Open'], [161698, 'Eyes Closed'], [169132, 'Eyes Open'], [170098, 'Eyes Closed'], [173600, 'Paused']], 'class_type': 'Non-vascular MCI', 'class_label': 1}}

----------------------------------------------------------------------------------------------------

torch.Size([20, 800])
{'signal': tensor([[-54., -56., -58.,  ...,  87.,  81.,  68.],
        [  5.,   5.,   2.,  ...,  39.,  34.,  26.],
        [  4.,   2.,   1.,  ...,   8.,   7.,   6.],
        ...,
        [ -6.,  -6.,  -3.,  ...,  11.,  13.,  14.],
        [  6.,   6.,   7.,  ...,  -6.,  -4.,   2.],
        [-66., -53., -22.,  ...,   7.,   5., -35.]]), 'age': tensor(-0.3053), 'class_label': tensor(0), 'metadata': {'serial': '00941', 'edfname': '01163965_130718', 'birth': '1950-08-06', 'record': '2018-07-13T14:16:08', 'age': 67, 'dx1': 'smi', 'label': ['normal', 'smi'], 'events': [[0, 'Start Recording'], [0, 'New Montage - Montage 002'], [36340, 'Eyes Open'], [71918, 'Eyes Closed'], [72926, 'Eyes Open'], [74062, 'Eyes Closed'], [78556, 'Eyes Open'], [80278, 'Eyes Closed'], [86982, 'Eyes Open'], [87746, 'Eyes Closed'], [89064, 'Photic On - 3.0 Hz'], [89272, 'Eyes Open'], [90406, 'Eyes Closed'], [91080, 'Photic Off'], [93138, 'Photic On - 6.0 Hz'], [95154, 'Photic Off'], [95446, 'Eyes Open'], [96538, 'Eyes Closed'], [97170, 'Photic On - 9.0 Hz'], [99228, 'Photic Off'], [101244, 'Photic On - 12.0 Hz'], [103259, 'Photic Off'], [103426, 'Eyes Open'], [104644, 'Eyes Closed'], [105318, 'Photic On - 15.0 Hz'], [107335, 'Photic Off'], [109352, 'Photic On - 18.0 Hz'], [111409, 'Photic Off'], [111872, 'Eyes Open'], [113048, 'Eyes Closed'], [113426, 'Photic On - 21.0 Hz'], [115442, 'Photic Off'], [117500, 'Photic On - 24.0 Hz'], [119516, 'Photic Off'], [121658, 'Photic On - 27.0 Hz'], [123590, 'Photic Off'], [125606, 'Photic On - 30.0 Hz'], [127622, 'Photic Off'], [127958, 'Eyes Open'], [129302, 'Eyes Closed'], [136316, 'Eyes Open'], [137492, 'Eyes Closed'], [141902, 'Eyes Open'], [143540, 'Eyes Closed'], [147068, 'Eyes Open'], [148286, 'Eyes Closed'], [156770, 'Eyes Open'], [158114, 'Eyes Closed'], [177854, 'Eyes Open'], [179324, 'Eyes Closed'], [181400, 'Paused']], 'class_type': 'Normal', 'class_label': 0}}

Data loader test

print('Current PyTorch device:', device)
if device.type == 'cuda':
    num_workers = 0 # A number other than 0 causes an error
    pin_memory = True
else:
    num_workers = 0
    pin_memory = False

train_loader = DataLoader(train_dataset, 
                          batch_size=32, 
                          shuffle=True, 
                          drop_last=True,
                          num_workers=num_workers, 
                          pin_memory=pin_memory,
                          collate_fn=eeg_collate_fn)

for i_batch, sample_batched in enumerate(train_loader):
    sample_batched['signal'].to(device)
    sample_batched['age'].to(device)
    sample_batched['class_label'].to(device)
    
    print(i_batch, 
          sample_batched['signal'].shape, 
          sample_batched['age'].shape, 
          sample_batched['class_label'].shape, 
          len(sample_batched['metadata']))
    
    if i_batch > 3:
        break

Current PyTorch device: cuda
0 torch.Size([32, 20, 800]) torch.Size([32]) torch.Size([32]) 32
1 torch.Size([32, 20, 800]) torch.Size([32]) torch.Size([32]) 32
2 torch.Size([32, 20, 800]) torch.Size([32]) torch.Size([32]) 32
3 torch.Size([32, 20, 800]) torch.Size([32]) torch.Size([32]) 32
4 torch.Size([32, 20, 800]) torch.Size([32]) torch.Size([32]) 32

Train, validation, test dataloaders

train_loader = DataLoader(train_dataset, 
                          batch_size=32, 
                          shuffle=True, 
                          drop_last=True,
                          num_workers=num_workers, 
                          pin_memory=pin_memory,
                          collate_fn=eeg_collate_fn)

val_loader = DataLoader(val_dataset, 
                        batch_size=32, 
                        shuffle=False, 
                        drop_last=False,
                        num_workers=num_workers, 
                        pin_memory=pin_memory,
                        collate_fn=eeg_collate_fn)

test_loader = DataLoader(test_dataset, 
                         batch_size=32, 
                         shuffle=False, 
                         drop_last=False,
                         num_workers=num_workers, 
                         pin_memory=pin_memory,
                         collate_fn=eeg_collate_fn)

---

Some useful functions for training

def count_parameters(model):
    return sum(p.numel() for p in model.parameters() if p.requires_grad)


def visualize_network_tensorboard(model, name):
    from torch.utils.tensorboard import SummaryWriter
    import ipynbname
    nb_fname = ipynbname.name()

    # default `log_dir` is "runs" - we'll be more specific here
    writer = SummaryWriter('runs/' + nb_fname + '_' + name)

    for batch_i, sample_batched in enumerate(train_loader):
        # pull up the batch data
        x = sample_batched['signal'].to(device)
        age = sample_batched['age'].to(device)
        target = sample_batched['class_label'].to(device)

        # apply model on whole batch directly on device
        writer.add_graph(model, (x, age))
        # output = model(x, age)
        break
        
    writer.close()

def train_one_epoch(model, optimizer, log_interval):
    # turn the models to training mode
    model.train()
    losses = []
    correct, total = (0, 0)
    
    C = len(class_label_to_type)
    train_confusion = np.zeros((C, C), dtype=np.int32)
    
    for batch_i, sample_batched in enumerate(train_loader):
        # pull up the batch data
        x = sample_batched['signal'].to(device)
        age = sample_batched['age'].to(device)
        target = sample_batched['class_label'].to(device)
        
        # apply model on whole batch directly on device
        output = model(x, age)

        # negative log-likelihood for a tensor of size (batch x n_output)
        pred = F.log_softmax(output, dim=1)
        loss = F.nll_loss(pred, target)
        
        # backprop and update
        loss.backward()
        optimizer.step()
        optimizer.zero_grad()
        
        # record loss
        losses.append(loss.item())

        # train accuracy
        pred = pred.argmax(dim=-1)
        correct += pred.squeeze().eq(target).sum().item()
        total += pred.shape[0]
        
        # confusion matrix
        train_confusion += calculate_confusion_matrix(pred, target)

        # print training stats
        if log_interval is not None and (batch_i + 1) % log_interval == 0:
            print(f'- Iter {batch_i + 1:03d} / {len(train_loader):03d}, Loss: {loss.item():.06f}')
    
    train_accuracy = 100.0 * correct / total
    return (losses, train_accuracy, train_confusion)


def check_val_accuracy(model, repeat=1):
    model.eval()
    correct, total = (0, 0)
        
    C = len(class_label_to_type)
    val_confusion = np.zeros((C, C), dtype=np.int32)
    
    for k in range(repeat):
        for sample_batched in val_loader:
            # pull up the data
            x = sample_batched['signal'].to(device)
            age = sample_batched['age'].to(device)
            target = sample_batched['class_label'].to(device)

            # apply model on whole batch directly on device
            output = model(x, age)
            pred = F.log_softmax(output, dim=1)

            # val accuracy
            pred = pred.argmax(dim=-1)
            correct += pred.squeeze().eq(target).sum().item()
            total += pred.shape[0]
            
            # confusion matrix
            val_confusion += calculate_confusion_matrix(pred, target)
            
    val_accuracy = 100.0 * correct / total
    return (val_accuracy, val_confusion)

    
def check_test_accuracy(model, repeat=1):
    model.eval()
    correct, total = (0, 0)
    
    C = len(class_label_to_type)
    test_confusion = np.zeros((C, C), dtype=np.int32)
    
    test_debug = {data['metadata']['serial']: 
                  {'GT': data['class_label'].item(), 
                   'Acc': 0, 
                   'Pred': [0] * C} for data in test_dataset}
        
    for k in range(repeat):
        for sample_batched in test_loader:
            # pull up the data
            x = sample_batched['signal'].to(device)
            age = sample_batched['age'].to(device)
            target = sample_batched['class_label'].to(device)

            # apply model on whole batch directly on device
            output = model(x, age)
            pred = F.log_softmax(output, dim=1)
            
            # test accuracy
            pred = pred.argmax(dim=-1)
            correct += pred.squeeze().eq(target).sum().item()
            total += pred.shape[0]
            
            # confusion matrix
            test_confusion += calculate_confusion_matrix(pred, target)
            
            # test debug
            for n in range(pred.shape[0]):
                serial = sample_batched['metadata'][n]['serial']
                test_debug[serial]['Pred'][pred[n].item()] += 1
                acc = test_debug[serial]['Pred'][target[n].item()] / np.sum(test_debug[serial]['Pred']) * 100
                test_debug[serial]['Acc'] = f'{acc:>6.02f}%'
        
    test_accuracy = 100.0 * correct / total
    return (test_accuracy, test_confusion, test_debug)



def calculate_confusion_matrix(pred, target):
    N = target.shape[0]
    C = len(class_label_to_type)
    confusion = np.zeros((C, C), dtype=np.int32)
    
    for i in range(N):
        r = target[i]
        c = pred[i]
        confusion[r, c] += 1
    return confusion

def draw_loss_plot(loss_history):
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic
    
    fig = plt.figure(num=1, clear=True, figsize=(15.0, 6.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    ax.plot(loss_history)
    
    ax.vlines(0, 0, 1, transform=ax.get_xaxis_transform(), colors='k', alpha=0.1)    
    for e in range(1, n_epoch + 1):
        
        if e % lr_schedule_step == 0:
            ax.vlines(e*len(train_loader) - 1, 0, 1, transform=ax.get_xaxis_transform(), colors='m', alpha=0.3)
        else:
            ax.vlines(e*len(train_loader) - 1, 0, 1, transform=ax.get_xaxis_transform(), colors='k', alpha=0.1)
    
    ax.set_title('Loss Plot')
    ax.set_xlabel('Iteration')
    ax.set_ylabel('Training Loss')
    
    plt.show()
    fig.clear()
    plt.close(fig)


def draw_accuracy_history(train_acc_history, val_acc_history):
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic
    
    fig = plt.figure(num=1, clear=True, figsize=(15.0, 6.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    ax.plot(train_acc_history, 'r-', label='Train accuracy')
    ax.plot(val_acc_history, 'b-', label='Validation accuracy')
    ax.legend(loc='lower right')
    
    ax.set_title('Accuracy Plot during Training')
    ax.set_xlabel('Epoch')
    ax.set_ylabel('Accuracy (%)')
    
    plt.show()
    fig.clear()
    plt.close(fig)


def draw_confusion(confusion):
    C = len(class_label_to_type)
    
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic
    plt.rcParams['image.cmap'] = 'jet' # 'nipy_spectral'

    fig = plt.figure(num=1, clear=True, figsize=(5.0, 5.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    im = ax.imshow(confusion, alpha=0.8)

    ax.set_xticks(np.arange(C))
    ax.set_yticks(np.arange(C))
    ax.set_xticklabels(class_label_to_type)
    ax.set_yticklabels(class_label_to_type)
    
    for r in range(C):
        for c in range(C):
            text = ax.text(c, r, confusion[r, c],
                           ha="center", va="center", color='k')
    
    ax.set_title('Confusion Matrix')
    ax.set_xlabel('Prediction')
    ax.set_ylabel('Ground Truth')
    plt.setp(ax.get_xticklabels(), rotation=45, ha="right", rotation_mode="anchor")
    
    plt.show()
    fig.clear()
    plt.close(fig)

def learning_rate_search(module, min_log_lr, max_log_lr, trials, epochs):
    learning_rate_record = []
    for t in tqdm(range(trials)):
        log_lr = np.random.uniform(min_log_lr, max_log_lr)
        lr = 10 ** log_lr
        
        module.reset_weights()
        optimizer = optim.AdamW(model.parameters(), lr=lr, weight_decay=0.0001)

        for e in range(epochs):
            _, train_accuracy, _ = train_one_epoch(model, optimizer, log_interval=None)
        
        # Train accuracy for the final epoch is stored
        learning_rate_record.append((log_lr, train_accuracy))
    
    return learning_rate_record


def draw_learning_rate_record(learning_rate_record):
    plt.style.use('default') # default, ggplot, fivethirtyeight, classic

    fig = plt.figure(num=1, clear=True, figsize=(8.0, 8.0), constrained_layout=True)
    ax = fig.add_subplot(1, 1, 1)
    
    ax.set_title('Learning Rate Search')
    ax.set_xlabel('Learning rate in log-scale')
    ax.set_ylabel('Train accuracy')
    
    for log_lr, val_accuracy in learning_rate_record:
        ax.scatter(log_lr, val_accuracy, c='r',
                   alpha=0.5, edgecolors='none')
    
    plt.show()
    fig.clear()
    plt.close(fig)

---

Train 1D CNN models

1D CNN: Siamese model

class SiameseNet(nn.Module):
    def __init__(self, n_input=20, n_output=3, n_channel=64,
                 use_age=True, final_pool='average'):
        super().__init__()
        
        if final_pool not in {'average', 'max'}:
            raise ValueError("final_pool must be set to one of ['average', 'max']")
        
        self.use_age = use_age
        
        self.conv1 = nn.Conv1d(n_input, n_channel, kernel_size=25, dilation=2, stride=3)
        self.bn1 = nn.BatchNorm1d(n_channel)
        self.pool1 = nn.MaxPool1d(3)
        
        self.conv2 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn2 = nn.BatchNorm1d(n_channel)
        self.pool2 = nn.MaxPool1d(2)
        
        self.conv3 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn3 = nn.BatchNorm1d(n_channel)
        self.pool3 = nn.MaxPool1d(2)
        
        self.conv4_1 = nn.Conv1d(n_channel, n_channel*2, kernel_size=5)
        self.bn4_1 = nn.BatchNorm1d(n_channel*2)

        self.conv4_2 = nn.Conv1d(n_channel, n_channel*2, kernel_size=5)
        self.bn4_2 = nn.BatchNorm1d(n_channel*2)

        if final_pool == 'average':
            self.final_pool = nn.AdaptiveAvgPool1d(1)
        elif final_pool == 'max':
            self.final_pool = nn.AdaptiveMaxPool1d(1)
            
        if self.use_age:        
            self.fc1 = nn.Linear(n_channel*2 + 1, n_channel)
        else:
            self.fc1 = nn.Linear(n_channel*2, n_channel)
            
        self.dropout = nn.Dropout(p=0.3)
        self.bnfc1 = nn.BatchNorm1d(n_channel)
        self.fc2 = nn.Linear(n_channel, n_output)
        
    def reset_weights(self):
        for m in self.modules():
            if hasattr(m, 'reset_parameters'):
                m.reset_parameters()

    def forward(self, x, age):
        # conv-bn-relu-pool 
        x = self.conv1(x)
        x = F.relu(self.bn1(x))
        x = self.pool1(x)

        x = self.conv2(x)
        x = F.relu(self.bn2(x))
        x = self.pool2(x)

        x = self.conv3(x)
        x = F.relu(self.bn3(x))
        x = self.pool3(x)
        
        L = x.shape[2]
        x1 = x[:, :, :L//2]
        x2 = x[:, :, L//2:]
        
        x1 = self.conv4_1(x1)
        x1 = F.relu(self.bn4_1(x1))

        x2 = self.conv4_2(x2)
        x2 = F.relu(self.bn4_2(x2))
        
        x1 = self.final_pool(x1).squeeze()
        x2 = self.final_pool(x2).squeeze()

        x = x1 * x2
        x = x.reshape(x.shape[0], -1) # (N, C, 1) -> (N, C)
        if self.use_age:
            x = torch.cat((x, age.reshape(-1, 1)), dim=1)

        # fc-bn-dropout-relu-fc
        x = self.fc1(x)
        x = self.bnfc1(x)
        x = self.dropout(x)
        x = F.relu(x)
        x = self.fc2(x)
        
        return x
        # return F.log_softmax(x, dim=1)

model = SiameseNet(n_input=train_dataset[0]['signal'].shape[0], 
                n_output=3, 
                use_age=True, 
                final_pool='average')

model = model.to(device, dtype=torch.float32)
print(model)
print()

# tensorboard visualization
visualize_network_tensorboard(model, 'SiameseNet_1')

# number of parameters
n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

SiameseNet(
  (conv1): Conv1d(20, 64, kernel_size=(25,), stride=(3,), dilation=(2,))
  (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool1): MaxPool1d(kernel_size=3, stride=3, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv3): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn3): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool3): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv4_1): Conv1d(64, 128, kernel_size=(5,), stride=(1,))
  (bn4_1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv4_2): Conv1d(64, 128, kernel_size=(5,), stride=(1,))
  (bn4_2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (final_pool): AdaptiveAvgPool1d(output_size=1)
  (fc1): Linear(in_features=129, out_features=64, bias=True)
  (dropout): Dropout(p=0.3, inplace=False)
  (bnfc1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc2): Linear(in_features=64, out_features=3, bias=True)
)

C:\Users\IPIS-Minjae\anaconda3\envs\EEG_Project\lib\site-packages\torch\nn\functional.py:652: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at  ..\c10/core/TensorImpl.h:1156.)
  return torch.max_pool1d(input, kernel_size, stride, padding, dilation, ceil_mode)
C:\Users\IPIS-Minjae\anaconda3\envs\EEG_Project\lib\site-packages\torch\_tensor.py:575: UserWarning: floor_divide is deprecated, and will be removed in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values.
To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor'). (Triggered internally at  ..\aten\src\ATen\native\BinaryOps.cpp:467.)
  return torch.floor_divide(self, other)

The Number of parameters of the model: 181,251

# record = learning_rate_search(model, 
#                               min_log_lr=-4.5, 
#                               max_log_lr=-1.0, 
#                               trials=500, 
#                               epochs=1)
# draw_learning_rate_record(record)
# best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

best_log_lr = -1.8528783104568398
print('best_log_lr:', best_log_lr)

best_log_lr: -1.8528783104568398

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 200
log_interval = len(train_loader) // 2

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d}, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e} {"-"*30}')
    
    # train
    loss, train_accuracy, train_confusion = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy, val_confusion = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train accuracy / confusion: {train_accuracy:.2f}% / {train_confusion.tolist()}, ')
    print(f'* Val   accuracy / confusion: {val_accuracy:.2f}% / {val_confusion.tolist()}')
    print()
    
    
# test
test_accuracy, test_confusion, test_debug = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print()
print('- Confusion matrix:\n', test_confusion)
print()
print('- Debug table:')
pprint.pp(test_debug, indent=4)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(test_confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.364191
- Iter 024 / 025, Loss: 1.010918

* Train accuracy / confusion: 46.00% / [[256, 86, 14], [152, 105, 13], [98, 69, 7]], 
* Val   accuracy / confusion: 50.00% / [[39, 3, 4], [19, 9, 7], [15, 4, 4]]

------------------------------ Epoch 002 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.106244
- Iter 024 / 025, Loss: 0.991897

* Train accuracy / confusion: 48.25% / [[287, 61, 11], [166, 84, 16], [102, 58, 15]], 
* Val   accuracy / confusion: 46.15% / [[44, 2, 0], [31, 4, 0], [21, 2, 0]]

------------------------------ Epoch 003 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.987058
- Iter 024 / 025, Loss: 1.073795

* Train accuracy / confusion: 50.00% / [[262, 76, 16], [137, 107, 24], [70, 77, 31]], 
* Val   accuracy / confusion: 55.77% / [[34, 12, 0], [11, 24, 0], [9, 14, 0]]

------------------------------ Epoch 004 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.973178
- Iter 024 / 025, Loss: 1.050247

* Train accuracy / confusion: 50.88% / [[270, 76, 8], [126, 126, 18], [68, 97, 11]], 
* Val   accuracy / confusion: 52.88% / [[33, 12, 1], [10, 22, 3], [9, 14, 0]]

------------------------------ Epoch 005 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.073630
- Iter 024 / 025, Loss: 1.102940

* Train accuracy / confusion: 51.00% / [[263, 78, 14], [132, 111, 26], [59, 83, 34]], 
* Val   accuracy / confusion: 59.62% / [[34, 12, 0], [9, 26, 0], [7, 14, 2]]

------------------------------ Epoch 006 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.149424
- Iter 024 / 025, Loss: 1.091944

* Train accuracy / confusion: 52.50% / [[296, 50, 10], [143, 107, 18], [71, 88, 17]], 
* Val   accuracy / confusion: 52.88% / [[33, 13, 0], [13, 21, 1], [7, 15, 1]]

------------------------------ Epoch 007 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.883473
- Iter 024 / 025, Loss: 0.767687

* Train accuracy / confusion: 50.12% / [[265, 80, 9], [131, 115, 24], [68, 87, 21]], 
* Val   accuracy / confusion: 51.92% / [[39, 7, 0], [20, 15, 0], [17, 6, 0]]

------------------------------ Epoch 008 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.154236
- Iter 024 / 025, Loss: 0.976737

* Train accuracy / confusion: 51.38% / [[272, 76, 10], [127, 125, 18], [68, 90, 14]], 
* Val   accuracy / confusion: 54.81% / [[36, 10, 0], [14, 21, 0], [10, 13, 0]]

------------------------------ Epoch 009 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.967032
- Iter 024 / 025, Loss: 0.920471

* Train accuracy / confusion: 54.50% / [[300, 32, 24], [147, 80, 43], [77, 41, 56]], 
* Val   accuracy / confusion: 52.88% / [[35, 11, 0], [14, 20, 1], [10, 13, 0]]

------------------------------ Epoch 010 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.006643
- Iter 024 / 025, Loss: 0.950611

* Train accuracy / confusion: 51.00% / [[268, 75, 17], [136, 114, 16], [64, 84, 26]], 
* Val   accuracy / confusion: 55.77% / [[37, 9, 0], [14, 20, 1], [11, 11, 1]]

------------------------------ Epoch 011 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.936156
- Iter 024 / 025, Loss: 1.149983

* Train accuracy / confusion: 50.62% / [[284, 58, 17], [141, 85, 39], [73, 67, 36]], 
* Val   accuracy / confusion: 55.77% / [[34, 12, 0], [11, 24, 0], [8, 15, 0]]

------------------------------ Epoch 012 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.044630
- Iter 024 / 025, Loss: 0.937397

* Train accuracy / confusion: 53.50% / [[273, 74, 9], [120, 125, 20], [62, 87, 30]], 
* Val   accuracy / confusion: 54.81% / [[31, 14, 1], [8, 24, 3], [5, 16, 2]]

------------------------------ Epoch 013 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.019846
- Iter 024 / 025, Loss: 0.839384

* Train accuracy / confusion: 53.00% / [[246, 101, 6], [99, 156, 16], [52, 102, 22]], 
* Val   accuracy / confusion: 52.88% / [[33, 13, 0], [13, 21, 1], [7, 15, 1]]

------------------------------ Epoch 014 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.101805
- Iter 024 / 025, Loss: 0.951784

* Train accuracy / confusion: 52.75% / [[262, 89, 5], [114, 141, 12], [62, 96, 19]], 
* Val   accuracy / confusion: 54.81% / [[34, 11, 1], [10, 21, 4], [6, 15, 2]]

------------------------------ Epoch 015 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.990758
- Iter 024 / 025, Loss: 1.012663

* Train accuracy / confusion: 54.62% / [[277, 66, 16], [118, 119, 35], [50, 78, 41]], 
* Val   accuracy / confusion: 52.88% / [[33, 12, 1], [11, 21, 3], [6, 16, 1]]

------------------------------ Epoch 016 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.896310
- Iter 024 / 025, Loss: 1.023859

* Train accuracy / confusion: 53.12% / [[259, 79, 18], [112, 139, 16], [57, 93, 27]], 
* Val   accuracy / confusion: 56.73% / [[33, 13, 0], [8, 26, 1], [6, 17, 0]]

------------------------------ Epoch 017 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.252245
- Iter 024 / 025, Loss: 0.918857

* Train accuracy / confusion: 52.38% / [[264, 80, 14], [119, 125, 25], [63, 80, 30]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [9, 25, 1], [6, 17, 0]]

------------------------------ Epoch 018 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.840521
- Iter 024 / 025, Loss: 0.907667

* Train accuracy / confusion: 49.62% / [[264, 79, 13], [139, 104, 26], [63, 83, 29]], 
* Val   accuracy / confusion: 56.73% / [[33, 12, 1], [8, 25, 2], [5, 17, 1]]

------------------------------ Epoch 019 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.060644
- Iter 024 / 025, Loss: 0.878364

* Train accuracy / confusion: 52.50% / [[259, 81, 10], [118, 135, 23], [55, 93, 26]], 
* Val   accuracy / confusion: 53.85% / [[35, 9, 2], [11, 20, 4], [7, 15, 1]]

------------------------------ Epoch 020 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.925347
- Iter 024 / 025, Loss: 1.087862

* Train accuracy / confusion: 54.12% / [[267, 79, 6], [120, 133, 19], [55, 88, 33]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [9, 19, 7], [6, 15, 2]]

------------------------------ Epoch 021 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.137245
- Iter 024 / 025, Loss: 1.168267

* Train accuracy / confusion: 51.88% / [[261, 80, 17], [113, 122, 36], [55, 84, 32]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [10, 22, 3], [6, 16, 1]]

------------------------------ Epoch 022 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.158338
- Iter 024 / 025, Loss: 1.040200

* Train accuracy / confusion: 53.12% / [[276, 63, 20], [126, 100, 41], [63, 62, 49]], 
* Val   accuracy / confusion: 54.81% / [[34, 11, 1], [9, 20, 6], [7, 13, 3]]

------------------------------ Epoch 023 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.222152
- Iter 024 / 025, Loss: 0.973001

* Train accuracy / confusion: 55.00% / [[262, 79, 14], [110, 141, 20], [51, 86, 37]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [9, 20, 6], [6, 15, 2]]

------------------------------ Epoch 024 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.991867
- Iter 024 / 025, Loss: 0.936585

* Train accuracy / confusion: 53.25% / [[261, 80, 14], [117, 130, 22], [62, 79, 35]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [9, 20, 6], [6, 15, 2]]

------------------------------ Epoch 025 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.931037
- Iter 024 / 025, Loss: 0.994023

* Train accuracy / confusion: 54.00% / [[273, 68, 13], [122, 121, 27], [57, 81, 38]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [8, 23, 4], [6, 16, 1]]

------------------------------ Epoch 026 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.985838
- Iter 024 / 025, Loss: 0.856397

* Train accuracy / confusion: 54.62% / [[272, 67, 20], [115, 113, 39], [53, 69, 52]], 
* Val   accuracy / confusion: 55.77% / [[32, 12, 2], [9, 23, 3], [7, 13, 3]]

------------------------------ Epoch 027 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.891761
- Iter 024 / 025, Loss: 1.040871

* Train accuracy / confusion: 54.25% / [[263, 80, 10], [109, 129, 34], [51, 82, 42]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [7, 24, 4], [5, 17, 1]]

------------------------------ Epoch 028 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.975463
- Iter 024 / 025, Loss: 1.024818

* Train accuracy / confusion: 54.25% / [[272, 72, 10], [111, 128, 28], [55, 90, 34]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 029 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.955088
- Iter 024 / 025, Loss: 0.862638

* Train accuracy / confusion: 54.50% / [[255, 89, 11], [103, 144, 20], [57, 84, 37]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [7, 22, 6], [8, 12, 3]]

------------------------------ Epoch 030 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.959726
- Iter 024 / 025, Loss: 0.874602

* Train accuracy / confusion: 54.38% / [[271, 70, 17], [115, 123, 31], [59, 73, 41]], 
* Val   accuracy / confusion: 55.77% / [[34, 11, 1], [10, 21, 4], [6, 14, 3]]

------------------------------ Epoch 031 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.993723
- Iter 024 / 025, Loss: 0.908746

* Train accuracy / confusion: 54.50% / [[265, 78, 11], [110, 137, 20], [56, 89, 34]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [7, 23, 5], [6, 15, 2]]

------------------------------ Epoch 032 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.945620
- Iter 024 / 025, Loss: 0.853816

* Train accuracy / confusion: 53.12% / [[257, 86, 12], [112, 125, 31], [56, 78, 43]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [9, 21, 5], [5, 15, 3]]

------------------------------ Epoch 033 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.819115
- Iter 024 / 025, Loss: 0.780871

* Train accuracy / confusion: 53.38% / [[259, 75, 18], [121, 114, 33], [61, 65, 54]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [7, 22, 6], [5, 16, 2]]

------------------------------ Epoch 034 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.918784
- Iter 024 / 025, Loss: 0.862584

* Train accuracy / confusion: 54.00% / [[264, 81, 15], [111, 131, 27], [58, 76, 37]], 
* Val   accuracy / confusion: 50.96% / [[35, 7, 4], [12, 15, 8], [9, 11, 3]]

------------------------------ Epoch 035 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.896001
- Iter 024 / 025, Loss: 1.081337

* Train accuracy / confusion: 52.75% / [[270, 63, 19], [128, 97, 45], [60, 63, 55]], 
* Val   accuracy / confusion: 56.73% / [[33, 11, 2], [6, 23, 6], [6, 14, 3]]

------------------------------ Epoch 036 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.075453
- Iter 024 / 025, Loss: 1.093352

* Train accuracy / confusion: 55.75% / [[271, 72, 14], [105, 135, 29], [58, 76, 40]], 
* Val   accuracy / confusion: 60.58% / [[34, 11, 1], [7, 24, 4], [10, 8, 5]]

------------------------------ Epoch 037 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.777710
- Iter 024 / 025, Loss: 1.067545

* Train accuracy / confusion: 54.75% / [[267, 80, 8], [114, 143, 12], [55, 93, 28]], 
* Val   accuracy / confusion: 54.81% / [[34, 10, 2], [10, 18, 7], [6, 12, 5]]

------------------------------ Epoch 038 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.818107
- Iter 024 / 025, Loss: 0.994466

* Train accuracy / confusion: 55.88% / [[276, 67, 16], [114, 122, 31], [59, 66, 49]], 
* Val   accuracy / confusion: 52.88% / [[35, 8, 3], [11, 18, 6], [8, 13, 2]]

------------------------------ Epoch 039 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.963631
- Iter 024 / 025, Loss: 1.193090

* Train accuracy / confusion: 53.50% / [[260, 80, 12], [114, 122, 33], [54, 79, 46]], 
* Val   accuracy / confusion: 53.85% / [[34, 11, 1], [10, 21, 4], [8, 14, 1]]

------------------------------ Epoch 040 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.952856
- Iter 024 / 025, Loss: 0.877497

* Train accuracy / confusion: 54.12% / [[254, 95, 6], [108, 142, 21], [55, 82, 37]], 
* Val   accuracy / confusion: 50.00% / [[32, 13, 1], [9, 19, 7], [5, 17, 1]]

------------------------------ Epoch 041 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.104259
- Iter 024 / 025, Loss: 0.832767

* Train accuracy / confusion: 54.50% / [[263, 83, 11], [109, 133, 27], [55, 79, 40]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [8, 20, 7], [6, 16, 1]]

------------------------------ Epoch 042 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.010795
- Iter 024 / 025, Loss: 1.246474

* Train accuracy / confusion: 53.88% / [[254, 89, 15], [94, 134, 37], [52, 82, 43]], 
* Val   accuracy / confusion: 52.88% / [[31, 14, 1], [7, 22, 6], [5, 16, 2]]

------------------------------ Epoch 043 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.990896
- Iter 024 / 025, Loss: 0.909166

* Train accuracy / confusion: 53.62% / [[264, 78, 11], [121, 131, 19], [57, 85, 34]], 
* Val   accuracy / confusion: 53.85% / [[35, 9, 2], [9, 20, 6], [6, 16, 1]]

------------------------------ Epoch 044 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.981015
- Iter 024 / 025, Loss: 1.049241

* Train accuracy / confusion: 55.62% / [[276, 61, 17], [120, 113, 39], [61, 57, 56]], 
* Val   accuracy / confusion: 55.77% / [[34, 11, 1], [11, 23, 1], [7, 15, 1]]

------------------------------ Epoch 045 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.946060
- Iter 024 / 025, Loss: 0.922448

* Train accuracy / confusion: 53.00% / [[262, 90, 8], [109, 136, 19], [58, 92, 26]], 
* Val   accuracy / confusion: 54.81% / [[35, 8, 3], [11, 19, 5], [9, 11, 3]]

------------------------------ Epoch 046 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.890034
- Iter 024 / 025, Loss: 0.985564

* Train accuracy / confusion: 53.38% / [[258, 85, 13], [129, 111, 28], [58, 60, 58]], 
* Val   accuracy / confusion: 50.00% / [[34, 9, 3], [10, 17, 8], [5, 17, 1]]

------------------------------ Epoch 047 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.860637
- Iter 024 / 025, Loss: 1.232266

* Train accuracy / confusion: 54.75% / [[263, 82, 9], [111, 135, 23], [60, 77, 40]], 
* Val   accuracy / confusion: 53.85% / [[34, 10, 2], [10, 21, 4], [5, 17, 1]]

------------------------------ Epoch 048 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.834636
- Iter 024 / 025, Loss: 1.061218

* Train accuracy / confusion: 55.75% / [[271, 68, 19], [107, 129, 32], [53, 75, 46]], 
* Val   accuracy / confusion: 55.77% / [[36, 10, 0], [13, 20, 2], [6, 15, 2]]

------------------------------ Epoch 049 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.913089
- Iter 024 / 025, Loss: 0.942474

* Train accuracy / confusion: 55.50% / [[286, 62, 13], [115, 125, 23], [59, 84, 33]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [9, 20, 6], [7, 12, 4]]

------------------------------ Epoch 050 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.020765
- Iter 024 / 025, Loss: 0.917420

* Train accuracy / confusion: 54.75% / [[261, 79, 16], [116, 129, 22], [54, 75, 48]], 
* Val   accuracy / confusion: 54.81% / [[35, 9, 2], [12, 19, 4], [8, 12, 3]]

------------------------------ Epoch 051 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.822392
- Iter 024 / 025, Loss: 1.035932

* Train accuracy / confusion: 52.88% / [[268, 70, 17], [127, 103, 38], [59, 66, 52]], 
* Val   accuracy / confusion: 53.85% / [[31, 13, 2], [7, 24, 4], [5, 17, 1]]

------------------------------ Epoch 052 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.896064
- Iter 024 / 025, Loss: 0.944591

* Train accuracy / confusion: 54.75% / [[261, 85, 12], [106, 125, 33], [55, 71, 52]], 
* Val   accuracy / confusion: 50.00% / [[32, 11, 3], [11, 18, 6], [8, 13, 2]]

------------------------------ Epoch 053 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.893274
- Iter 024 / 025, Loss: 0.984684

* Train accuracy / confusion: 55.50% / [[258, 82, 15], [105, 146, 19], [49, 86, 40]], 
* Val   accuracy / confusion: 50.96% / [[31, 11, 4], [5, 18, 12], [5, 14, 4]]

------------------------------ Epoch 054 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.971517
- Iter 024 / 025, Loss: 0.967886

* Train accuracy / confusion: 55.62% / [[259, 75, 22], [106, 136, 28], [54, 70, 50]], 
* Val   accuracy / confusion: 59.62% / [[35, 9, 2], [6, 25, 4], [6, 15, 2]]

------------------------------ Epoch 055 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.866559
- Iter 024 / 025, Loss: 0.952277

* Train accuracy / confusion: 54.88% / [[252, 98, 9], [96, 142, 27], [52, 79, 45]], 
* Val   accuracy / confusion: 53.85% / [[34, 10, 2], [8, 21, 6], [8, 14, 1]]

------------------------------ Epoch 056 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.977632
- Iter 024 / 025, Loss: 0.875027

* Train accuracy / confusion: 54.50% / [[262, 75, 14], [116, 121, 33], [60, 66, 53]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 057 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.917604
- Iter 024 / 025, Loss: 0.901230

* Train accuracy / confusion: 57.12% / [[267, 80, 9], [104, 144, 20], [52, 78, 46]], 
* Val   accuracy / confusion: 58.65% / [[34, 10, 2], [8, 21, 6], [7, 10, 6]]

------------------------------ Epoch 058 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.767443
- Iter 024 / 025, Loss: 0.987053

* Train accuracy / confusion: 55.50% / [[260, 76, 18], [112, 133, 23], [59, 68, 51]], 
* Val   accuracy / confusion: 53.85% / [[33, 8, 5], [8, 18, 9], [11, 7, 5]]

------------------------------ Epoch 059 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.866295
- Iter 024 / 025, Loss: 0.914887

* Train accuracy / confusion: 55.88% / [[269, 71, 17], [118, 118, 30], [65, 52, 60]], 
* Val   accuracy / confusion: 52.88% / [[34, 10, 2], [10, 17, 8], [5, 14, 4]]

------------------------------ Epoch 060 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.942729
- Iter 024 / 025, Loss: 0.858156

* Train accuracy / confusion: 54.75% / [[262, 89, 6], [103, 144, 19], [55, 90, 32]], 
* Val   accuracy / confusion: 49.04% / [[33, 12, 1], [9, 17, 9], [5, 17, 1]]

------------------------------ Epoch 061 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.010759
- Iter 024 / 025, Loss: 1.003522

* Train accuracy / confusion: 56.38% / [[268, 74, 15], [105, 126, 37], [53, 65, 57]], 
* Val   accuracy / confusion: 53.85% / [[34, 11, 1], [9, 20, 6], [6, 15, 2]]

------------------------------ Epoch 062 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.989153
- Iter 024 / 025, Loss: 0.820756

* Train accuracy / confusion: 55.00% / [[249, 92, 16], [92, 149, 29], [54, 77, 42]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [6, 24, 5], [6, 15, 2]]

------------------------------ Epoch 063 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.997972
- Iter 024 / 025, Loss: 0.805238

* Train accuracy / confusion: 55.50% / [[264, 75, 17], [117, 136, 20], [60, 67, 44]], 
* Val   accuracy / confusion: 50.96% / [[29, 15, 2], [7, 22, 6], [7, 14, 2]]

------------------------------ Epoch 064 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.882383
- Iter 024 / 025, Loss: 0.888963

* Train accuracy / confusion: 55.50% / [[260, 88, 9], [107, 133, 29], [54, 69, 51]], 
* Val   accuracy / confusion: 52.88% / [[34, 11, 1], [10, 19, 6], [6, 15, 2]]

------------------------------ Epoch 065 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.051582
- Iter 024 / 025, Loss: 1.016720

* Train accuracy / confusion: 55.75% / [[259, 88, 11], [104, 143, 22], [50, 79, 44]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [7, 22, 6], [9, 12, 2]]

------------------------------ Epoch 066 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.102567
- Iter 024 / 025, Loss: 0.929680

* Train accuracy / confusion: 54.25% / [[254, 85, 16], [114, 125, 31], [53, 67, 55]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [7, 20, 8], [6, 13, 4]]

------------------------------ Epoch 067 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.032147
- Iter 024 / 025, Loss: 0.851081

* Train accuracy / confusion: 54.62% / [[257, 85, 15], [111, 134, 24], [56, 72, 46]], 
* Val   accuracy / confusion: 55.77% / [[32, 14, 0], [7, 22, 6], [6, 13, 4]]

------------------------------ Epoch 068 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.901637
- Iter 024 / 025, Loss: 0.897257

* Train accuracy / confusion: 56.88% / [[258, 89, 16], [87, 149, 28], [57, 68, 48]], 
* Val   accuracy / confusion: 50.96% / [[34, 9, 3], [9, 19, 7], [5, 18, 0]]

------------------------------ Epoch 069 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.975043
- Iter 024 / 025, Loss: 0.980532

* Train accuracy / confusion: 55.50% / [[255, 94, 10], [108, 118, 38], [50, 56, 71]], 
* Val   accuracy / confusion: 50.00% / [[30, 12, 4], [5, 19, 11], [3, 17, 3]]

------------------------------ Epoch 070 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.119939
- Iter 024 / 025, Loss: 1.010646

* Train accuracy / confusion: 53.25% / [[271, 71, 16], [122, 105, 35], [63, 67, 50]], 
* Val   accuracy / confusion: 54.81% / [[32, 13, 1], [8, 24, 3], [7, 15, 1]]

------------------------------ Epoch 071 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.053221
- Iter 024 / 025, Loss: 0.933261

* Train accuracy / confusion: 53.88% / [[268, 83, 4], [117, 131, 18], [57, 90, 32]], 
* Val   accuracy / confusion: 54.81% / [[35, 7, 4], [11, 17, 7], [8, 10, 5]]

------------------------------ Epoch 072 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.828932
- Iter 024 / 025, Loss: 1.015461

* Train accuracy / confusion: 55.00% / [[268, 69, 21], [112, 109, 49], [56, 53, 63]], 
* Val   accuracy / confusion: 55.77% / [[35, 8, 3], [9, 17, 9], [8, 9, 6]]

------------------------------ Epoch 073 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.906729
- Iter 024 / 025, Loss: 0.991118

* Train accuracy / confusion: 54.12% / [[260, 86, 4], [119, 131, 23], [65, 70, 42]], 
* Val   accuracy / confusion: 55.77% / [[37, 9, 0], [13, 19, 3], [11, 10, 2]]

------------------------------ Epoch 074 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.876283
- Iter 024 / 025, Loss: 0.955158

* Train accuracy / confusion: 57.12% / [[270, 71, 14], [113, 121, 36], [59, 50, 66]], 
* Val   accuracy / confusion: 49.04% / [[28, 15, 3], [6, 20, 9], [5, 15, 3]]

------------------------------ Epoch 075 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.746431
- Iter 024 / 025, Loss: 1.163300

* Train accuracy / confusion: 55.88% / [[255, 91, 10], [92, 146, 28], [51, 81, 46]], 
* Val   accuracy / confusion: 52.88% / [[35, 8, 3], [10, 17, 8], [8, 12, 3]]

------------------------------ Epoch 076 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.844144
- Iter 024 / 025, Loss: 1.237792

* Train accuracy / confusion: 56.00% / [[250, 96, 13], [86, 162, 20], [50, 87, 36]], 
* Val   accuracy / confusion: 55.77% / [[35, 9, 2], [10, 21, 4], [7, 14, 2]]

------------------------------ Epoch 077 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.979007
- Iter 024 / 025, Loss: 0.847287

* Train accuracy / confusion: 56.62% / [[272, 64, 16], [117, 120, 35], [53, 62, 61]], 
* Val   accuracy / confusion: 50.96% / [[31, 8, 7], [8, 18, 9], [6, 13, 4]]

------------------------------ Epoch 078 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.938860
- Iter 024 / 025, Loss: 0.989486

* Train accuracy / confusion: 54.38% / [[247, 93, 12], [110, 139, 25], [48, 77, 49]], 
* Val   accuracy / confusion: 56.73% / [[31, 14, 1], [7, 24, 4], [4, 15, 4]]

------------------------------ Epoch 079 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.751111
- Iter 024 / 025, Loss: 0.962366

* Train accuracy / confusion: 53.88% / [[257, 84, 17], [121, 126, 23], [58, 66, 48]], 
* Val   accuracy / confusion: 47.12% / [[34, 9, 3], [14, 13, 8], [12, 9, 2]]

------------------------------ Epoch 080 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.832297
- Iter 024 / 025, Loss: 0.840645

* Train accuracy / confusion: 55.25% / [[274, 66, 13], [127, 115, 28], [64, 60, 53]], 
* Val   accuracy / confusion: 50.96% / [[33, 12, 1], [12, 17, 6], [7, 13, 3]]

------------------------------ Epoch 081 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.873433
- Iter 024 / 025, Loss: 0.875505

* Train accuracy / confusion: 54.50% / [[251, 98, 8], [106, 133, 27], [55, 70, 52]], 
* Val   accuracy / confusion: 51.92% / [[30, 15, 1], [5, 22, 8], [6, 15, 2]]

------------------------------ Epoch 082 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.013882
- Iter 024 / 025, Loss: 1.037020

* Train accuracy / confusion: 54.38% / [[285, 58, 15], [143, 95, 32], [60, 57, 55]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [11, 18, 6], [6, 12, 5]]

------------------------------ Epoch 083 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.830871
- Iter 024 / 025, Loss: 0.999338

* Train accuracy / confusion: 55.62% / [[259, 84, 13], [95, 143, 28], [59, 76, 43]], 
* Val   accuracy / confusion: 50.96% / [[30, 15, 1], [10, 22, 3], [7, 15, 1]]

------------------------------ Epoch 084 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.192852
- Iter 024 / 025, Loss: 0.852903

* Train accuracy / confusion: 54.75% / [[247, 91, 15], [98, 138, 33], [56, 69, 53]], 
* Val   accuracy / confusion: 60.58% / [[36, 8, 2], [6, 23, 6], [6, 13, 4]]

------------------------------ Epoch 085 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.828487
- Iter 024 / 025, Loss: 0.924899

* Train accuracy / confusion: 54.88% / [[273, 66, 17], [130, 117, 21], [67, 60, 49]], 
* Val   accuracy / confusion: 50.96% / [[34, 8, 4], [11, 15, 9], [6, 13, 4]]

------------------------------ Epoch 086 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.898173
- Iter 024 / 025, Loss: 1.054795

* Train accuracy / confusion: 54.50% / [[267, 76, 14], [125, 108, 36], [59, 54, 61]], 
* Val   accuracy / confusion: 50.00% / [[35, 7, 4], [11, 12, 12], [8, 10, 5]]

------------------------------ Epoch 087 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.053928
- Iter 024 / 025, Loss: 0.907733

* Train accuracy / confusion: 55.75% / [[268, 77, 13], [117, 111, 36], [57, 54, 67]], 
* Val   accuracy / confusion: 53.85% / [[32, 13, 1], [11, 20, 4], [10, 9, 4]]

------------------------------ Epoch 088 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.840808
- Iter 024 / 025, Loss: 0.885524

* Train accuracy / confusion: 55.38% / [[273, 76, 9], [122, 123, 27], [54, 69, 47]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [5, 25, 5], [6, 16, 1]]

------------------------------ Epoch 089 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.878056
- Iter 024 / 025, Loss: 0.886498

* Train accuracy / confusion: 57.12% / [[252, 91, 15], [89, 142, 32], [58, 58, 63]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [6, 22, 7], [10, 10, 3]]

------------------------------ Epoch 090 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.012297
- Iter 024 / 025, Loss: 0.796913

* Train accuracy / confusion: 54.25% / [[262, 91, 6], [112, 124, 32], [58, 67, 48]], 
* Val   accuracy / confusion: 50.96% / [[26, 18, 2], [4, 26, 5], [7, 15, 1]]

------------------------------ Epoch 091 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.065889
- Iter 024 / 025, Loss: 1.013097

* Train accuracy / confusion: 54.12% / [[262, 82, 11], [121, 129, 22], [63, 68, 42]], 
* Val   accuracy / confusion: 59.62% / [[36, 6, 4], [7, 21, 7], [7, 11, 5]]

------------------------------ Epoch 092 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.973842
- Iter 024 / 025, Loss: 0.818245

* Train accuracy / confusion: 57.12% / [[244, 96, 14], [86, 151, 35], [50, 62, 62]], 
* Val   accuracy / confusion: 51.92% / [[27, 18, 1], [3, 26, 6], [3, 19, 1]]

------------------------------ Epoch 093 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.840372
- Iter 024 / 025, Loss: 0.965324

* Train accuracy / confusion: 56.75% / [[267, 75, 11], [115, 133, 21], [63, 61, 54]], 
* Val   accuracy / confusion: 55.77% / [[34, 8, 4], [7, 19, 9], [5, 13, 5]]

------------------------------ Epoch 094 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.828375
- Iter 024 / 025, Loss: 0.951705

* Train accuracy / confusion: 57.38% / [[290, 48, 20], [117, 114, 37], [65, 54, 55]], 
* Val   accuracy / confusion: 51.92% / [[30, 13, 3], [6, 21, 8], [6, 14, 3]]

------------------------------ Epoch 095 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.734167
- Iter 024 / 025, Loss: 1.002206

* Train accuracy / confusion: 56.25% / [[261, 83, 13], [104, 137, 28], [54, 68, 52]], 
* Val   accuracy / confusion: 57.69% / [[34, 11, 1], [8, 24, 3], [9, 12, 2]]

------------------------------ Epoch 096 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.008396
- Iter 024 / 025, Loss: 1.015310

* Train accuracy / confusion: 55.38% / [[265, 81, 13], [107, 133, 27], [51, 78, 45]], 
* Val   accuracy / confusion: 50.00% / [[33, 10, 3], [12, 14, 9], [7, 11, 5]]

------------------------------ Epoch 097 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.838957
- Iter 024 / 025, Loss: 0.979403

* Train accuracy / confusion: 57.00% / [[277, 75, 6], [110, 135, 22], [60, 71, 44]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [7, 21, 7], [7, 14, 2]]

------------------------------ Epoch 098 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.871050
- Iter 024 / 025, Loss: 0.868816

* Train accuracy / confusion: 56.12% / [[265, 82, 10], [103, 136, 30], [56, 70, 48]], 
* Val   accuracy / confusion: 51.92% / [[34, 10, 2], [10, 18, 7], [12, 9, 2]]

------------------------------ Epoch 099 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.913274
- Iter 024 / 025, Loss: 1.060498

* Train accuracy / confusion: 56.38% / [[269, 72, 15], [112, 120, 36], [56, 58, 62]], 
* Val   accuracy / confusion: 57.69% / [[36, 6, 4], [6, 20, 9], [7, 12, 4]]

------------------------------ Epoch 100 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.921572
- Iter 024 / 025, Loss: 0.771876

* Train accuracy / confusion: 55.50% / [[296, 54, 13], [139, 102, 24], [66, 60, 46]], 
* Val   accuracy / confusion: 54.81% / [[37, 7, 2], [11, 17, 7], [8, 12, 3]]

------------------------------ Epoch 101 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.876671
- Iter 024 / 025, Loss: 0.833648

* Train accuracy / confusion: 54.62% / [[267, 75, 17], [128, 108, 29], [54, 60, 62]], 
* Val   accuracy / confusion: 56.73% / [[29, 15, 2], [2, 28, 5], [3, 18, 2]]

------------------------------ Epoch 102 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.921660
- Iter 024 / 025, Loss: 0.822851

* Train accuracy / confusion: 58.00% / [[261, 85, 14], [100, 135, 32], [51, 54, 68]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [7, 17, 11], [9, 6, 8]]

------------------------------ Epoch 103 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.936744
- Iter 024 / 025, Loss: 0.780139

* Train accuracy / confusion: 56.25% / [[256, 74, 22], [105, 133, 32], [57, 60, 61]], 
* Val   accuracy / confusion: 56.73% / [[31, 14, 1], [4, 25, 6], [5, 15, 3]]

------------------------------ Epoch 104 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.858211
- Iter 024 / 025, Loss: 0.922531

* Train accuracy / confusion: 56.12% / [[260, 85, 6], [104, 144, 23], [58, 75, 45]], 
* Val   accuracy / confusion: 56.73% / [[34, 9, 3], [7, 20, 8], [6, 12, 5]]

------------------------------ Epoch 105 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.857848
- Iter 024 / 025, Loss: 1.015563

* Train accuracy / confusion: 57.75% / [[253, 84, 19], [94, 146, 31], [42, 68, 63]], 
* Val   accuracy / confusion: 58.65% / [[35, 10, 1], [8, 24, 3], [9, 12, 2]]

------------------------------ Epoch 106 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.924371
- Iter 024 / 025, Loss: 0.912593

* Train accuracy / confusion: 53.75% / [[285, 59, 11], [150, 94, 27], [70, 53, 51]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [6, 22, 7], [5, 13, 5]]

------------------------------ Epoch 107 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.054823
- Iter 024 / 025, Loss: 0.935291

* Train accuracy / confusion: 58.62% / [[241, 100, 13], [75, 168, 27], [50, 66, 60]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [4, 24, 7], [5, 15, 3]]

------------------------------ Epoch 108 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.003868
- Iter 024 / 025, Loss: 0.954005

* Train accuracy / confusion: 56.12% / [[245, 96, 12], [84, 158, 28], [49, 82, 46]], 
* Val   accuracy / confusion: 50.00% / [[29, 14, 3], [4, 22, 9], [4, 18, 1]]

------------------------------ Epoch 109 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.967689
- Iter 024 / 025, Loss: 0.841566

* Train accuracy / confusion: 56.75% / [[272, 68, 17], [101, 123, 42], [53, 65, 59]], 
* Val   accuracy / confusion: 51.92% / [[33, 9, 4], [10, 18, 7], [7, 13, 3]]

------------------------------ Epoch 110 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.999042
- Iter 024 / 025, Loss: 0.820925

* Train accuracy / confusion: 55.75% / [[265, 82, 9], [107, 137, 22], [59, 75, 44]], 
* Val   accuracy / confusion: 49.04% / [[37, 4, 5], [16, 10, 9], [11, 8, 4]]

------------------------------ Epoch 111 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.854018
- Iter 024 / 025, Loss: 0.732085

* Train accuracy / confusion: 57.00% / [[283, 61, 14], [131, 109, 27], [62, 49, 64]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [9, 18, 8], [3, 16, 4]]

------------------------------ Epoch 112 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.228835
- Iter 024 / 025, Loss: 0.888570

* Train accuracy / confusion: 53.75% / [[269, 72, 16], [117, 114, 35], [63, 67, 47]], 
* Val   accuracy / confusion: 59.62% / [[33, 12, 1], [7, 24, 4], [6, 12, 5]]

------------------------------ Epoch 113 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.124946
- Iter 024 / 025, Loss: 0.901523

* Train accuracy / confusion: 58.12% / [[268, 85, 8], [87, 153, 26], [61, 68, 44]], 
* Val   accuracy / confusion: 56.73% / [[35, 10, 1], [8, 21, 6], [8, 12, 3]]

------------------------------ Epoch 114 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.978669
- Iter 024 / 025, Loss: 1.062798

* Train accuracy / confusion: 57.62% / [[281, 56, 18], [111, 128, 30], [67, 57, 52]], 
* Val   accuracy / confusion: 58.65% / [[34, 8, 4], [7, 19, 9], [6, 9, 8]]

------------------------------ Epoch 115 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.972746
- Iter 024 / 025, Loss: 0.951856

* Train accuracy / confusion: 57.88% / [[262, 80, 12], [104, 144, 24], [57, 60, 57]], 
* Val   accuracy / confusion: 50.96% / [[31, 12, 3], [7, 19, 9], [9, 11, 3]]

------------------------------ Epoch 116 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.863535
- Iter 024 / 025, Loss: 0.899970

* Train accuracy / confusion: 57.00% / [[272, 73, 16], [97, 138, 30], [57, 71, 46]], 
* Val   accuracy / confusion: 53.85% / [[34, 11, 1], [12, 20, 3], [7, 14, 2]]

------------------------------ Epoch 117 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.000340
- Iter 024 / 025, Loss: 0.963746

* Train accuracy / confusion: 56.38% / [[273, 63, 18], [125, 114, 28], [69, 46, 64]], 
* Val   accuracy / confusion: 57.69% / [[33, 9, 4], [6, 22, 7], [4, 14, 5]]

------------------------------ Epoch 118 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.846490
- Iter 024 / 025, Loss: 0.842130

* Train accuracy / confusion: 56.50% / [[268, 73, 19], [107, 118, 38], [61, 50, 66]], 
* Val   accuracy / confusion: 45.19% / [[34, 9, 3], [13, 10, 12], [9, 11, 3]]

------------------------------ Epoch 119 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.848871
- Iter 024 / 025, Loss: 0.974842

* Train accuracy / confusion: 58.12% / [[275, 69, 9], [114, 138, 20], [59, 64, 52]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [7, 24, 4], [6, 14, 3]]

------------------------------ Epoch 120 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.062791
- Iter 024 / 025, Loss: 0.983631

* Train accuracy / confusion: 56.75% / [[278, 58, 19], [120, 118, 28], [58, 63, 58]], 
* Val   accuracy / confusion: 52.88% / [[36, 8, 2], [12, 18, 5], [7, 15, 1]]

------------------------------ Epoch 121 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.904620
- Iter 024 / 025, Loss: 0.927868

* Train accuracy / confusion: 57.88% / [[261, 84, 13], [98, 158, 13], [57, 72, 44]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [8, 21, 6], [6, 13, 4]]

------------------------------ Epoch 122 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.027726
- Iter 024 / 025, Loss: 0.734252

* Train accuracy / confusion: 56.88% / [[296, 55, 8], [120, 122, 26], [76, 60, 37]], 
* Val   accuracy / confusion: 59.62% / [[36, 9, 1], [8, 21, 6], [6, 12, 5]]

------------------------------ Epoch 123 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.783084
- Iter 024 / 025, Loss: 0.843125

* Train accuracy / confusion: 58.38% / [[273, 63, 21], [105, 139, 23], [60, 61, 55]], 
* Val   accuracy / confusion: 52.88% / [[27, 19, 0], [3, 25, 7], [3, 17, 3]]

------------------------------ Epoch 124 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.036278
- Iter 024 / 025, Loss: 0.862178

* Train accuracy / confusion: 58.75% / [[276, 66, 15], [112, 129, 29], [62, 46, 65]], 
* Val   accuracy / confusion: 59.62% / [[33, 10, 3], [4, 25, 6], [6, 13, 4]]

------------------------------ Epoch 125 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.761065
- Iter 024 / 025, Loss: 1.014052

* Train accuracy / confusion: 59.12% / [[259, 84, 11], [96, 150, 22], [57, 57, 64]], 
* Val   accuracy / confusion: 55.77% / [[39, 5, 2], [11, 15, 9], [10, 9, 4]]

------------------------------ Epoch 126 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.162525
- Iter 024 / 025, Loss: 1.001433

* Train accuracy / confusion: 56.00% / [[251, 84, 17], [101, 126, 44], [57, 49, 71]], 
* Val   accuracy / confusion: 50.00% / [[31, 13, 2], [10, 18, 7], [5, 15, 3]]

------------------------------ Epoch 127 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.791520
- Iter 024 / 025, Loss: 0.837415

* Train accuracy / confusion: 57.62% / [[297, 54, 7], [130, 125, 14], [63, 71, 39]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [7, 21, 7], [3, 17, 3]]

------------------------------ Epoch 128 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.103248
- Iter 024 / 025, Loss: 0.784078

* Train accuracy / confusion: 58.38% / [[256, 85, 11], [92, 151, 30], [48, 67, 60]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [11, 19, 5], [6, 14, 3]]

------------------------------ Epoch 129 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.910863
- Iter 024 / 025, Loss: 1.001266

* Train accuracy / confusion: 57.62% / [[273, 68, 14], [109, 126, 32], [57, 59, 62]], 
* Val   accuracy / confusion: 55.77% / [[30, 15, 1], [9, 23, 3], [5, 13, 5]]

------------------------------ Epoch 130 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.975393
- Iter 024 / 025, Loss: 0.899814

* Train accuracy / confusion: 57.12% / [[260, 76, 16], [106, 134, 30], [53, 62, 63]], 
* Val   accuracy / confusion: 50.00% / [[33, 10, 3], [12, 17, 6], [8, 13, 2]]

------------------------------ Epoch 131 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.850186
- Iter 024 / 025, Loss: 0.956434

* Train accuracy / confusion: 57.50% / [[277, 62, 15], [125, 120, 23], [63, 52, 63]], 
* Val   accuracy / confusion: 61.54% / [[37, 8, 1], [7, 23, 5], [7, 12, 4]]

------------------------------ Epoch 132 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.753889
- Iter 024 / 025, Loss: 0.931585

* Train accuracy / confusion: 58.25% / [[266, 76, 12], [106, 144, 18], [62, 60, 56]], 
* Val   accuracy / confusion: 52.88% / [[33, 12, 1], [9, 21, 5], [7, 15, 1]]

------------------------------ Epoch 133 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.904382
- Iter 024 / 025, Loss: 0.855990

* Train accuracy / confusion: 55.00% / [[253, 81, 20], [116, 119, 37], [53, 53, 68]], 
* Val   accuracy / confusion: 53.85% / [[35, 9, 2], [13, 18, 4], [8, 12, 3]]

------------------------------ Epoch 134 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.937973
- Iter 024 / 025, Loss: 1.001708

* Train accuracy / confusion: 58.12% / [[287, 65, 9], [111, 136, 22], [58, 70, 42]], 
* Val   accuracy / confusion: 56.73% / [[33, 12, 1], [7, 24, 4], [6, 15, 2]]

------------------------------ Epoch 135 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.893856
- Iter 024 / 025, Loss: 0.829789

* Train accuracy / confusion: 57.25% / [[283, 70, 6], [112, 129, 24], [68, 62, 46]], 
* Val   accuracy / confusion: 50.00% / [[32, 13, 1], [13, 17, 5], [9, 11, 3]]

------------------------------ Epoch 136 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.972738
- Iter 024 / 025, Loss: 0.838284

* Train accuracy / confusion: 56.12% / [[283, 65, 9], [122, 127, 18], [68, 69, 39]], 
* Val   accuracy / confusion: 53.85% / [[32, 13, 1], [7, 22, 6], [7, 14, 2]]

------------------------------ Epoch 137 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.791917
- Iter 024 / 025, Loss: 0.883669

* Train accuracy / confusion: 56.00% / [[271, 64, 18], [109, 118, 43], [60, 58, 59]], 
* Val   accuracy / confusion: 51.92% / [[27, 18, 1], [6, 23, 6], [3, 16, 4]]

------------------------------ Epoch 138 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.022544
- Iter 024 / 025, Loss: 0.860304

* Train accuracy / confusion: 58.38% / [[264, 80, 12], [93, 158, 21], [48, 79, 45]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [6, 20, 9], [4, 16, 3]]

------------------------------ Epoch 139 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.705930
- Iter 024 / 025, Loss: 1.102823

* Train accuracy / confusion: 57.25% / [[272, 67, 15], [107, 136, 26], [66, 61, 50]], 
* Val   accuracy / confusion: 55.77% / [[35, 9, 2], [8, 21, 6], [8, 13, 2]]

------------------------------ Epoch 140 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.787320
- Iter 024 / 025, Loss: 0.814308

* Train accuracy / confusion: 59.00% / [[268, 71, 17], [105, 134, 31], [53, 51, 70]], 
* Val   accuracy / confusion: 52.88% / [[31, 13, 2], [6, 20, 9], [6, 13, 4]]

------------------------------ Epoch 141 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.748016
- Iter 024 / 025, Loss: 0.883116

* Train accuracy / confusion: 58.38% / [[266, 79, 12], [93, 138, 36], [54, 59, 63]], 
* Val   accuracy / confusion: 57.69% / [[31, 13, 2], [2, 27, 6], [5, 16, 2]]

------------------------------ Epoch 142 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.885548
- Iter 024 / 025, Loss: 0.932065

* Train accuracy / confusion: 54.88% / [[277, 69, 8], [127, 122, 21], [63, 73, 40]], 
* Val   accuracy / confusion: 48.08% / [[28, 15, 3], [7, 20, 8], [6, 15, 2]]

------------------------------ Epoch 143 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.905917
- Iter 024 / 025, Loss: 0.818035

* Train accuracy / confusion: 55.12% / [[262, 86, 8], [112, 125, 32], [52, 69, 54]], 
* Val   accuracy / confusion: 46.15% / [[32, 11, 3], [16, 14, 5], [8, 13, 2]]

------------------------------ Epoch 144 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.725659
- Iter 024 / 025, Loss: 0.836599

* Train accuracy / confusion: 56.75% / [[283, 60, 13], [117, 116, 33], [59, 64, 55]], 
* Val   accuracy / confusion: 53.85% / [[31, 13, 2], [7, 21, 7], [4, 15, 4]]

------------------------------ Epoch 145 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.792766
- Iter 024 / 025, Loss: 0.789829

* Train accuracy / confusion: 58.88% / [[273, 62, 17], [107, 137, 28], [62, 53, 61]], 
* Val   accuracy / confusion: 48.08% / [[27, 17, 2], [9, 18, 8], [3, 15, 5]]

------------------------------ Epoch 146 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.766178
- Iter 024 / 025, Loss: 0.852600

* Train accuracy / confusion: 55.75% / [[250, 96, 9], [88, 156, 24], [60, 77, 40]], 
* Val   accuracy / confusion: 53.85% / [[32, 9, 5], [7, 19, 9], [6, 12, 5]]

------------------------------ Epoch 147 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.886085
- Iter 024 / 025, Loss: 0.819523

* Train accuracy / confusion: 59.88% / [[275, 66, 14], [96, 145, 29], [58, 58, 59]], 
* Val   accuracy / confusion: 53.85% / [[30, 11, 5], [9, 22, 4], [6, 13, 4]]

------------------------------ Epoch 148 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.872866
- Iter 024 / 025, Loss: 1.202630

* Train accuracy / confusion: 61.12% / [[277, 70, 8], [101, 151, 17], [61, 54, 61]], 
* Val   accuracy / confusion: 51.92% / [[28, 16, 2], [6, 24, 5], [8, 13, 2]]

------------------------------ Epoch 149 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.741652
- Iter 024 / 025, Loss: 1.125211

* Train accuracy / confusion: 60.62% / [[274, 76, 11], [95, 153, 20], [54, 59, 58]], 
* Val   accuracy / confusion: 62.50% / [[35, 9, 2], [6, 26, 3], [7, 12, 4]]

------------------------------ Epoch 150 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.910491
- Iter 024 / 025, Loss: 0.952949

* Train accuracy / confusion: 58.00% / [[256, 84, 14], [92, 141, 38], [52, 56, 67]], 
* Val   accuracy / confusion: 57.69% / [[37, 8, 1], [13, 17, 5], [7, 10, 6]]

------------------------------ Epoch 151 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.926151
- Iter 024 / 025, Loss: 0.895654

* Train accuracy / confusion: 57.88% / [[288, 64, 6], [122, 135, 10], [67, 68, 40]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [9, 21, 5], [9, 11, 3]]

------------------------------ Epoch 152 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.843186
- Iter 024 / 025, Loss: 0.811638

* Train accuracy / confusion: 59.00% / [[261, 73, 24], [87, 142, 39], [57, 48, 69]], 
* Val   accuracy / confusion: 53.85% / [[32, 13, 1], [7, 22, 6], [6, 15, 2]]

------------------------------ Epoch 153 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.921841
- Iter 024 / 025, Loss: 1.071260

* Train accuracy / confusion: 58.88% / [[269, 87, 5], [101, 154, 12], [54, 70, 48]], 
* Val   accuracy / confusion: 54.81% / [[35, 6, 5], [11, 18, 6], [5, 14, 4]]

------------------------------ Epoch 154 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.675418
- Iter 024 / 025, Loss: 0.771055

* Train accuracy / confusion: 59.62% / [[289, 55, 11], [119, 114, 38], [64, 36, 74]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [8, 23, 4], [5, 14, 4]]

------------------------------ Epoch 155 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.980229
- Iter 024 / 025, Loss: 1.057399

* Train accuracy / confusion: 58.38% / [[270, 76, 9], [102, 146, 23], [53, 70, 51]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [9, 16, 10], [7, 9, 7]]

------------------------------ Epoch 156 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.022371
- Iter 024 / 025, Loss: 1.161200

* Train accuracy / confusion: 58.12% / [[281, 64, 12], [108, 119, 43], [59, 49, 65]], 
* Val   accuracy / confusion: 51.92% / [[35, 6, 5], [13, 13, 9], [9, 8, 6]]

------------------------------ Epoch 157 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.768244
- Iter 024 / 025, Loss: 1.109385

* Train accuracy / confusion: 61.00% / [[294, 52, 10], [117, 131, 19], [60, 54, 63]], 
* Val   accuracy / confusion: 58.65% / [[36, 7, 3], [11, 20, 4], [5, 13, 5]]

------------------------------ Epoch 158 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.790428
- Iter 024 / 025, Loss: 0.798088

* Train accuracy / confusion: 56.88% / [[272, 70, 13], [123, 119, 22], [57, 60, 64]], 
* Val   accuracy / confusion: 50.00% / [[34, 10, 2], [11, 15, 9], [7, 13, 3]]

------------------------------ Epoch 159 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.913059
- Iter 024 / 025, Loss: 0.998617

* Train accuracy / confusion: 59.62% / [[275, 63, 11], [94, 155, 26], [52, 77, 47]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [8, 22, 5], [6, 14, 3]]

------------------------------ Epoch 160 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.790512
- Iter 024 / 025, Loss: 0.957193

* Train accuracy / confusion: 62.25% / [[278, 62, 15], [82, 155, 31], [53, 59, 65]], 
* Val   accuracy / confusion: 57.69% / [[32, 14, 0], [6, 24, 5], [6, 13, 4]]

------------------------------ Epoch 161 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.833298
- Iter 024 / 025, Loss: 1.228294

* Train accuracy / confusion: 55.75% / [[252, 92, 10], [106, 144, 23], [61, 62, 50]], 
* Val   accuracy / confusion: 50.96% / [[37, 7, 2], [16, 13, 6], [10, 10, 3]]

------------------------------ Epoch 162 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.045859
- Iter 024 / 025, Loss: 1.012096

* Train accuracy / confusion: 59.88% / [[288, 51, 18], [102, 118, 47], [65, 38, 73]], 
* Val   accuracy / confusion: 56.73% / [[31, 15, 0], [7, 26, 2], [3, 18, 2]]

------------------------------ Epoch 163 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.682956
- Iter 024 / 025, Loss: 1.025492

* Train accuracy / confusion: 58.62% / [[289, 63, 4], [114, 147, 7], [65, 78, 33]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [9, 20, 6], [5, 13, 5]]

------------------------------ Epoch 164 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.851743
- Iter 024 / 025, Loss: 0.848189

* Train accuracy / confusion: 57.25% / [[264, 72, 18], [108, 131, 32], [56, 56, 63]], 
* Val   accuracy / confusion: 53.85% / [[36, 5, 5], [12, 17, 6], [6, 14, 3]]

------------------------------ Epoch 165 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.955746
- Iter 024 / 025, Loss: 0.801671

* Train accuracy / confusion: 60.12% / [[284, 60, 17], [111, 128, 27], [57, 47, 69]], 
* Val   accuracy / confusion: 57.69% / [[35, 10, 1], [8, 22, 5], [7, 13, 3]]

------------------------------ Epoch 166 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.957035
- Iter 024 / 025, Loss: 0.707237

* Train accuracy / confusion: 56.88% / [[278, 62, 14], [114, 129, 27], [64, 64, 48]], 
* Val   accuracy / confusion: 52.88% / [[33, 13, 0], [11, 18, 6], [5, 14, 4]]

------------------------------ Epoch 167 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.890078
- Iter 024 / 025, Loss: 1.126957

* Train accuracy / confusion: 58.38% / [[273, 68, 15], [99, 145, 24], [56, 71, 49]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [7, 22, 6], [6, 13, 4]]

------------------------------ Epoch 168 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.836329
- Iter 024 / 025, Loss: 0.914688

* Train accuracy / confusion: 57.50% / [[274, 68, 17], [121, 120, 25], [57, 52, 66]], 
* Val   accuracy / confusion: 60.58% / [[35, 8, 3], [9, 21, 5], [8, 8, 7]]

------------------------------ Epoch 169 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.844545
- Iter 024 / 025, Loss: 0.787595

* Train accuracy / confusion: 57.62% / [[290, 57, 11], [126, 110, 33], [60, 52, 61]], 
* Val   accuracy / confusion: 51.92% / [[31, 14, 1], [8, 21, 6], [6, 15, 2]]

------------------------------ Epoch 170 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.800254
- Iter 024 / 025, Loss: 0.769968

* Train accuracy / confusion: 59.75% / [[279, 62, 10], [107, 146, 18], [60, 65, 53]], 
* Val   accuracy / confusion: 56.73% / [[35, 9, 2], [11, 19, 5], [8, 10, 5]]

------------------------------ Epoch 171 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.754956
- Iter 024 / 025, Loss: 0.806364

* Train accuracy / confusion: 60.50% / [[256, 90, 14], [76, 165, 22], [49, 65, 63]], 
* Val   accuracy / confusion: 53.85% / [[35, 11, 0], [12, 17, 6], [6, 13, 4]]

------------------------------ Epoch 172 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.049070
- Iter 024 / 025, Loss: 0.937310

* Train accuracy / confusion: 58.38% / [[282, 61, 16], [111, 126, 28], [67, 50, 59]], 
* Val   accuracy / confusion: 54.81% / [[35, 7, 4], [13, 15, 7], [6, 10, 7]]

------------------------------ Epoch 173 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.874959
- Iter 024 / 025, Loss: 1.031054

* Train accuracy / confusion: 60.12% / [[272, 70, 9], [89, 164, 19], [56, 76, 45]], 
* Val   accuracy / confusion: 55.77% / [[35, 7, 4], [12, 17, 6], [7, 10, 6]]

------------------------------ Epoch 174 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.838154
- Iter 024 / 025, Loss: 0.959954

* Train accuracy / confusion: 60.88% / [[281, 53, 21], [101, 129, 38], [57, 43, 77]], 
* Val   accuracy / confusion: 59.62% / [[32, 9, 5], [7, 25, 3], [6, 12, 5]]

------------------------------ Epoch 175 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.940554
- Iter 024 / 025, Loss: 0.874666

* Train accuracy / confusion: 58.50% / [[257, 82, 13], [95, 156, 20], [52, 70, 55]], 
* Val   accuracy / confusion: 54.81% / [[36, 5, 5], [11, 17, 7], [8, 11, 4]]

------------------------------ Epoch 176 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 1.064598
- Iter 024 / 025, Loss: 1.005767

* Train accuracy / confusion: 59.38% / [[275, 64, 15], [101, 133, 36], [58, 51, 67]], 
* Val   accuracy / confusion: 55.77% / [[33, 12, 1], [10, 20, 5], [8, 10, 5]]

------------------------------ Epoch 177 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.980780
- Iter 024 / 025, Loss: 0.974962

* Train accuracy / confusion: 59.00% / [[288, 58, 9], [110, 148, 13], [67, 71, 36]], 
* Val   accuracy / confusion: 55.77% / [[32, 10, 4], [9, 22, 4], [6, 13, 4]]

------------------------------ Epoch 178 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.910671
- Iter 024 / 025, Loss: 1.165201

* Train accuracy / confusion: 59.50% / [[284, 55, 18], [111, 125, 31], [63, 46, 67]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [12, 19, 4], [9, 10, 4]]

------------------------------ Epoch 179 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.638109
- Iter 024 / 025, Loss: 0.858346

* Train accuracy / confusion: 57.75% / [[288, 58, 9], [121, 125, 23], [64, 63, 49]], 
* Val   accuracy / confusion: 50.96% / [[35, 9, 2], [13, 13, 9], [11, 7, 5]]

------------------------------ Epoch 180 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.887052
- Iter 024 / 025, Loss: 0.853831

* Train accuracy / confusion: 61.50% / [[280, 63, 15], [84, 149, 32], [48, 66, 63]], 
* Val   accuracy / confusion: 56.73% / [[31, 10, 5], [7, 22, 6], [3, 14, 6]]

------------------------------ Epoch 181 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.763253
- Iter 024 / 025, Loss: 0.810495

* Train accuracy / confusion: 59.50% / [[273, 61, 17], [91, 147, 34], [58, 63, 56]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [10, 18, 7], [7, 11, 5]]

------------------------------ Epoch 182 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.874937
- Iter 024 / 025, Loss: 0.865898

* Train accuracy / confusion: 61.50% / [[259, 81, 16], [80, 164, 23], [49, 59, 69]], 
* Val   accuracy / confusion: 52.88% / [[31, 13, 2], [10, 19, 6], [7, 11, 5]]

------------------------------ Epoch 183 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.862773
- Iter 024 / 025, Loss: 1.126885

* Train accuracy / confusion: 59.62% / [[270, 73, 13], [108, 131, 30], [45, 54, 76]], 
* Val   accuracy / confusion: 60.58% / [[34, 10, 2], [6, 24, 5], [6, 12, 5]]

------------------------------ Epoch 184 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.842494
- Iter 024 / 025, Loss: 0.683881

* Train accuracy / confusion: 59.75% / [[270, 72, 14], [94, 150, 24], [57, 61, 58]], 
* Val   accuracy / confusion: 55.77% / [[35, 9, 2], [10, 20, 5], [7, 13, 3]]

------------------------------ Epoch 185 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.927198
- Iter 024 / 025, Loss: 0.715049

* Train accuracy / confusion: 59.50% / [[283, 64, 12], [101, 137, 30], [52, 65, 56]], 
* Val   accuracy / confusion: 54.81% / [[30, 13, 3], [10, 20, 5], [8, 8, 7]]

------------------------------ Epoch 186 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.934411
- Iter 024 / 025, Loss: 0.754830

* Train accuracy / confusion: 61.38% / [[278, 69, 7], [92, 156, 23], [47, 71, 57]], 
* Val   accuracy / confusion: 55.77% / [[30, 15, 1], [7, 25, 3], [5, 15, 3]]

------------------------------ Epoch 187 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.848424
- Iter 024 / 025, Loss: 0.818854

* Train accuracy / confusion: 57.00% / [[250, 85, 17], [95, 142, 34], [51, 62, 64]], 
* Val   accuracy / confusion: 57.69% / [[34, 12, 0], [9, 24, 2], [8, 13, 2]]

------------------------------ Epoch 188 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.849669
- Iter 024 / 025, Loss: 0.952391

* Train accuracy / confusion: 61.62% / [[271, 78, 9], [72, 162, 31], [55, 62, 60]], 
* Val   accuracy / confusion: 56.73% / [[35, 7, 4], [9, 21, 5], [10, 10, 3]]

------------------------------ Epoch 189 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.822101
- Iter 024 / 025, Loss: 0.817448

* Train accuracy / confusion: 60.12% / [[276, 64, 12], [100, 140, 30], [55, 58, 65]], 
* Val   accuracy / confusion: 50.96% / [[26, 17, 3], [5, 23, 7], [7, 12, 4]]

------------------------------ Epoch 190 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.791098
- Iter 024 / 025, Loss: 0.961142

* Train accuracy / confusion: 62.25% / [[279, 66, 13], [81, 153, 32], [48, 62, 66]], 
* Val   accuracy / confusion: 54.81% / [[37, 6, 3], [14, 15, 6], [12, 6, 5]]

------------------------------ Epoch 191 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.786684
- Iter 024 / 025, Loss: 1.107522

* Train accuracy / confusion: 58.12% / [[282, 64, 11], [106, 130, 30], [56, 68, 53]], 
* Val   accuracy / confusion: 50.96% / [[34, 7, 5], [14, 17, 4], [9, 12, 2]]

------------------------------ Epoch 192 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.833363
- Iter 024 / 025, Loss: 0.804255

* Train accuracy / confusion: 61.50% / [[285, 52, 15], [105, 145, 24], [50, 62, 62]], 
* Val   accuracy / confusion: 56.73% / [[36, 8, 2], [9, 20, 6], [9, 11, 3]]

------------------------------ Epoch 193 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.832342
- Iter 024 / 025, Loss: 0.923484

* Train accuracy / confusion: 59.38% / [[270, 76, 10], [101, 138, 30], [54, 54, 67]], 
* Val   accuracy / confusion: 58.65% / [[31, 11, 4], [6, 24, 5], [8, 9, 6]]

------------------------------ Epoch 194 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.740891
- Iter 024 / 025, Loss: 1.003475

* Train accuracy / confusion: 61.00% / [[276, 67, 15], [99, 153, 17], [57, 57, 59]], 
* Val   accuracy / confusion: 57.69% / [[34, 10, 2], [12, 21, 2], [8, 10, 5]]

------------------------------ Epoch 195 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.742558
- Iter 024 / 025, Loss: 0.938650

* Train accuracy / confusion: 63.25% / [[289, 51, 13], [91, 149, 32], [56, 51, 68]], 
* Val   accuracy / confusion: 55.77% / [[35, 10, 1], [11, 19, 5], [11, 8, 4]]

------------------------------ Epoch 196 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.759121
- Iter 024 / 025, Loss: 0.828760

* Train accuracy / confusion: 61.88% / [[279, 63, 11], [82, 155, 33], [54, 62, 61]], 
* Val   accuracy / confusion: 47.12% / [[34, 9, 3], [15, 11, 9], [6, 13, 4]]

------------------------------ Epoch 197 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.967703
- Iter 024 / 025, Loss: 0.874678

* Train accuracy / confusion: 63.12% / [[285, 63, 9], [87, 151, 29], [51, 56, 69]], 
* Val   accuracy / confusion: 53.85% / [[36, 5, 5], [11, 16, 8], [7, 12, 4]]

------------------------------ Epoch 198 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.845905
- Iter 024 / 025, Loss: 0.828747

* Train accuracy / confusion: 59.62% / [[283, 66, 11], [111, 127, 28], [63, 44, 67]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [13, 17, 5], [8, 10, 5]]

------------------------------ Epoch 199 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.844657
- Iter 024 / 025, Loss: 0.915503

* Train accuracy / confusion: 59.88% / [[273, 70, 16], [96, 144, 28], [52, 59, 62]], 
* Val   accuracy / confusion: 54.81% / [[31, 10, 5], [10, 19, 6], [8, 8, 7]]

------------------------------ Epoch 200 / 500, Learning rate: 1.40e-02 ------------------------------
- Iter 012 / 025, Loss: 0.804006
- Iter 024 / 025, Loss: 0.794570

* Train accuracy / confusion: 61.50% / [[285, 60, 9], [93, 156, 22], [58, 66, 51]], 
* Val   accuracy / confusion: 59.62% / [[35, 10, 1], [5, 23, 7], [7, 12, 4]]

------------------------------ Epoch 201 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.832965
- Iter 024 / 025, Loss: 0.683665

* Train accuracy / confusion: 60.88% / [[282, 65, 17], [91, 141, 29], [52, 59, 64]], 
* Val   accuracy / confusion: 54.81% / [[31, 9, 6], [6, 21, 8], [6, 12, 5]]

------------------------------ Epoch 202 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.826556
- Iter 024 / 025, Loss: 0.929783

* Train accuracy / confusion: 64.00% / [[284, 58, 14], [85, 156, 25], [48, 58, 72]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [8, 21, 6], [4, 14, 5]]

------------------------------ Epoch 203 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.668355
- Iter 024 / 025, Loss: 0.941431

* Train accuracy / confusion: 59.88% / [[272, 72, 15], [88, 144, 35], [49, 62, 63]], 
* Val   accuracy / confusion: 57.69% / [[32, 12, 2], [8, 23, 4], [4, 14, 5]]

------------------------------ Epoch 204 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.804028
- Iter 024 / 025, Loss: 0.802438

* Train accuracy / confusion: 62.25% / [[274, 67, 14], [84, 157, 28], [62, 47, 67]], 
* Val   accuracy / confusion: 58.65% / [[32, 12, 2], [5, 24, 6], [7, 11, 5]]

------------------------------ Epoch 205 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.844481
- Iter 024 / 025, Loss: 0.760429

* Train accuracy / confusion: 62.12% / [[288, 58, 11], [100, 135, 29], [47, 58, 74]], 
* Val   accuracy / confusion: 56.73% / [[34, 10, 2], [8, 20, 7], [8, 10, 5]]

------------------------------ Epoch 206 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.880656
- Iter 024 / 025, Loss: 0.800647

* Train accuracy / confusion: 64.38% / [[286, 58, 12], [89, 153, 25], [43, 58, 76]], 
* Val   accuracy / confusion: 57.69% / [[34, 10, 2], [7, 22, 6], [7, 12, 4]]

------------------------------ Epoch 207 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.816134
- Iter 024 / 025, Loss: 1.118611

* Train accuracy / confusion: 62.62% / [[281, 63, 14], [88, 150, 30], [51, 53, 70]], 
* Val   accuracy / confusion: 50.00% / [[32, 8, 6], [9, 19, 7], [7, 15, 1]]

------------------------------ Epoch 208 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.842148
- Iter 024 / 025, Loss: 0.673460

* Train accuracy / confusion: 64.38% / [[275, 65, 12], [77, 168, 25], [59, 47, 72]], 
* Val   accuracy / confusion: 52.88% / [[31, 14, 1], [12, 19, 4], [4, 14, 5]]

------------------------------ Epoch 209 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.873959
- Iter 024 / 025, Loss: 0.782638

* Train accuracy / confusion: 64.38% / [[290, 55, 11], [86, 156, 25], [46, 62, 69]], 
* Val   accuracy / confusion: 61.54% / [[35, 10, 1], [5, 25, 5], [6, 13, 4]]

------------------------------ Epoch 210 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.801339
- Iter 024 / 025, Loss: 0.684685

* Train accuracy / confusion: 64.62% / [[284, 65, 10], [80, 159, 24], [59, 45, 74]], 
* Val   accuracy / confusion: 50.96% / [[29, 12, 5], [8, 21, 6], [11, 9, 3]]

------------------------------ Epoch 211 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.660425
- Iter 024 / 025, Loss: 0.907985

* Train accuracy / confusion: 61.38% / [[278, 65, 12], [94, 147, 26], [49, 63, 66]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [8, 20, 7], [7, 11, 5]]

------------------------------ Epoch 212 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.771034
- Iter 024 / 025, Loss: 0.679505

* Train accuracy / confusion: 64.00% / [[275, 64, 14], [83, 164, 24], [52, 51, 73]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [6, 18, 11], [9, 11, 3]]

------------------------------ Epoch 213 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.691413
- Iter 024 / 025, Loss: 0.710651

* Train accuracy / confusion: 63.25% / [[271, 74, 10], [89, 158, 24], [51, 46, 77]], 
* Val   accuracy / confusion: 55.77% / [[28, 16, 2], [6, 24, 5], [5, 12, 6]]

------------------------------ Epoch 214 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.868232
- Iter 024 / 025, Loss: 0.747948

* Train accuracy / confusion: 62.50% / [[280, 68, 13], [83, 153, 32], [46, 58, 67]], 
* Val   accuracy / confusion: 56.73% / [[34, 10, 2], [7, 21, 7], [6, 13, 4]]

------------------------------ Epoch 215 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.816797
- Iter 024 / 025, Loss: 0.948220

* Train accuracy / confusion: 65.00% / [[296, 48, 12], [90, 154, 23], [44, 63, 70]], 
* Val   accuracy / confusion: 58.65% / [[34, 11, 1], [6, 23, 6], [6, 13, 4]]

------------------------------ Epoch 216 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.669146
- Iter 024 / 025, Loss: 0.860119

* Train accuracy / confusion: 65.38% / [[288, 59, 10], [92, 162, 16], [50, 50, 73]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [9, 20, 6], [4, 14, 5]]

------------------------------ Epoch 217 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.962117
- Iter 024 / 025, Loss: 0.826976

* Train accuracy / confusion: 64.50% / [[287, 57, 14], [93, 156, 19], [55, 46, 73]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [8, 19, 8], [7, 11, 5]]

------------------------------ Epoch 218 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.890345
- Iter 024 / 025, Loss: 0.867714

* Train accuracy / confusion: 65.38% / [[287, 58, 16], [91, 157, 16], [44, 52, 79]], 
* Val   accuracy / confusion: 58.65% / [[34, 10, 2], [8, 20, 7], [7, 9, 7]]

------------------------------ Epoch 219 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.812014
- Iter 024 / 025, Loss: 0.802373

* Train accuracy / confusion: 64.50% / [[285, 59, 13], [88, 160, 22], [42, 60, 71]], 
* Val   accuracy / confusion: 51.92% / [[29, 16, 1], [9, 22, 4], [6, 14, 3]]

------------------------------ Epoch 220 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.898357
- Iter 024 / 025, Loss: 0.766606

* Train accuracy / confusion: 61.12% / [[285, 62, 10], [90, 143, 35], [49, 65, 61]], 
* Val   accuracy / confusion: 54.81% / [[34, 9, 3], [9, 20, 6], [9, 11, 3]]

------------------------------ Epoch 221 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.709263
- Iter 024 / 025, Loss: 0.710624

* Train accuracy / confusion: 64.12% / [[281, 53, 21], [95, 157, 20], [50, 48, 75]], 
* Val   accuracy / confusion: 55.77% / [[31, 13, 2], [6, 24, 5], [6, 14, 3]]

------------------------------ Epoch 222 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.918220
- Iter 024 / 025, Loss: 0.653274

* Train accuracy / confusion: 67.12% / [[289, 61, 7], [81, 164, 22], [47, 45, 84]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [10, 19, 6], [5, 14, 4]]

------------------------------ Epoch 223 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.825721
- Iter 024 / 025, Loss: 0.791491

* Train accuracy / confusion: 64.00% / [[290, 50, 17], [89, 153, 23], [47, 62, 69]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [10, 20, 5], [9, 10, 4]]

------------------------------ Epoch 224 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.778155
- Iter 024 / 025, Loss: 0.748378

* Train accuracy / confusion: 64.25% / [[292, 53, 14], [86, 153, 29], [54, 50, 69]], 
* Val   accuracy / confusion: 55.77% / [[34, 11, 1], [10, 19, 6], [6, 12, 5]]

------------------------------ Epoch 225 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.701840
- Iter 024 / 025, Loss: 0.766429

* Train accuracy / confusion: 63.62% / [[286, 60, 12], [83, 154, 29], [46, 61, 69]], 
* Val   accuracy / confusion: 51.92% / [[30, 12, 4], [9, 21, 5], [7, 13, 3]]

------------------------------ Epoch 226 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.775891
- Iter 024 / 025, Loss: 0.754983

* Train accuracy / confusion: 65.38% / [[293, 52, 14], [91, 154, 22], [47, 51, 76]], 
* Val   accuracy / confusion: 59.62% / [[32, 12, 2], [8, 23, 4], [4, 12, 7]]

------------------------------ Epoch 227 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.923627
- Iter 024 / 025, Loss: 0.874892

* Train accuracy / confusion: 63.88% / [[284, 59, 15], [89, 155, 22], [48, 56, 72]], 
* Val   accuracy / confusion: 59.62% / [[32, 13, 1], [9, 23, 3], [7, 9, 7]]

------------------------------ Epoch 228 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.708729
- Iter 024 / 025, Loss: 0.652937

* Train accuracy / confusion: 64.00% / [[283, 60, 15], [87, 162, 22], [48, 56, 67]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [9, 16, 10], [7, 7, 9]]

------------------------------ Epoch 229 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.815634
- Iter 024 / 025, Loss: 0.998299

* Train accuracy / confusion: 64.88% / [[289, 59, 7], [98, 146, 25], [57, 35, 84]], 
* Val   accuracy / confusion: 57.69% / [[33, 11, 2], [9, 21, 5], [6, 11, 6]]

------------------------------ Epoch 230 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 1.122153
- Iter 024 / 025, Loss: 0.760566

* Train accuracy / confusion: 65.00% / [[282, 62, 13], [84, 158, 23], [50, 48, 80]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [9, 21, 5], [8, 10, 5]]

------------------------------ Epoch 231 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.610619
- Iter 024 / 025, Loss: 0.814152

* Train accuracy / confusion: 63.75% / [[293, 51, 13], [94, 144, 31], [45, 56, 73]], 
* Val   accuracy / confusion: 55.77% / [[34, 8, 4], [10, 20, 5], [7, 12, 4]]

------------------------------ Epoch 232 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.614244
- Iter 024 / 025, Loss: 0.811806

* Train accuracy / confusion: 65.38% / [[291, 52, 14], [85, 159, 23], [42, 61, 73]], 
* Val   accuracy / confusion: 47.12% / [[28, 15, 3], [12, 15, 8], [5, 12, 6]]

------------------------------ Epoch 233 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.810964
- Iter 024 / 025, Loss: 0.764122

* Train accuracy / confusion: 65.50% / [[288, 55, 12], [83, 163, 22], [60, 44, 73]], 
* Val   accuracy / confusion: 55.77% / [[34, 12, 0], [11, 19, 5], [5, 13, 5]]

------------------------------ Epoch 234 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.956207
- Iter 024 / 025, Loss: 0.699448

* Train accuracy / confusion: 64.00% / [[290, 59, 8], [91, 151, 26], [51, 53, 71]], 
* Val   accuracy / confusion: 54.81% / [[30, 14, 2], [8, 22, 5], [7, 11, 5]]

------------------------------ Epoch 235 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.895597
- Iter 024 / 025, Loss: 0.782020

* Train accuracy / confusion: 64.50% / [[284, 58, 15], [82, 165, 21], [48, 60, 67]], 
* Val   accuracy / confusion: 57.69% / [[34, 9, 3], [12, 19, 4], [7, 9, 7]]

------------------------------ Epoch 236 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.765594
- Iter 024 / 025, Loss: 0.666888

* Train accuracy / confusion: 63.88% / [[289, 51, 15], [98, 143, 30], [45, 50, 79]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [10, 19, 6], [7, 11, 5]]

------------------------------ Epoch 237 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.925027
- Iter 024 / 025, Loss: 0.691911

* Train accuracy / confusion: 66.50% / [[294, 54, 9], [80, 159, 33], [43, 49, 79]], 
* Val   accuracy / confusion: 51.92% / [[34, 9, 3], [13, 14, 8], [11, 6, 6]]

------------------------------ Epoch 238 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.627056
- Iter 024 / 025, Loss: 0.805406

* Train accuracy / confusion: 65.88% / [[293, 52, 11], [87, 153, 27], [37, 59, 81]], 
* Val   accuracy / confusion: 63.46% / [[37, 9, 0], [10, 22, 3], [9, 7, 7]]

------------------------------ Epoch 239 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.771613
- Iter 024 / 025, Loss: 0.644360

* Train accuracy / confusion: 65.25% / [[289, 50, 12], [88, 154, 30], [51, 47, 79]], 
* Val   accuracy / confusion: 50.00% / [[34, 11, 1], [12, 14, 9], [8, 11, 4]]

------------------------------ Epoch 240 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.789728
- Iter 024 / 025, Loss: 0.779552

* Train accuracy / confusion: 65.00% / [[282, 55, 14], [90, 162, 22], [50, 49, 76]], 
* Val   accuracy / confusion: 47.12% / [[32, 12, 2], [15, 14, 6], [6, 14, 3]]

------------------------------ Epoch 241 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.661000
- Iter 024 / 025, Loss: 0.729330

* Train accuracy / confusion: 65.75% / [[297, 43, 13], [87, 160, 25], [55, 51, 69]], 
* Val   accuracy / confusion: 54.81% / [[33, 12, 1], [8, 21, 6], [8, 12, 3]]

------------------------------ Epoch 242 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.682351
- Iter 024 / 025, Loss: 0.933779

* Train accuracy / confusion: 65.38% / [[291, 53, 11], [91, 159, 18], [54, 50, 73]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [11, 21, 3], [6, 14, 3]]

------------------------------ Epoch 243 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.829358
- Iter 024 / 025, Loss: 0.699391

* Train accuracy / confusion: 63.88% / [[282, 67, 9], [88, 155, 24], [50, 51, 74]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [13, 19, 3], [7, 14, 2]]

------------------------------ Epoch 244 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.755840
- Iter 024 / 025, Loss: 0.741253

* Train accuracy / confusion: 66.50% / [[295, 54, 11], [86, 161, 24], [41, 52, 76]], 
* Val   accuracy / confusion: 50.00% / [[31, 13, 2], [13, 18, 4], [8, 12, 3]]

------------------------------ Epoch 245 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.872719
- Iter 024 / 025, Loss: 0.787560

* Train accuracy / confusion: 64.62% / [[287, 55, 10], [92, 150, 28], [50, 48, 80]], 
* Val   accuracy / confusion: 50.00% / [[31, 12, 3], [11, 19, 5], [7, 14, 2]]

------------------------------ Epoch 246 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.917801
- Iter 024 / 025, Loss: 0.852005

* Train accuracy / confusion: 64.88% / [[289, 53, 12], [85, 160, 27], [53, 51, 70]], 
* Val   accuracy / confusion: 54.81% / [[30, 13, 3], [10, 23, 2], [6, 13, 4]]

------------------------------ Epoch 247 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.719342
- Iter 024 / 025, Loss: 0.749923

* Train accuracy / confusion: 65.75% / [[292, 54, 8], [95, 150, 25], [44, 48, 84]], 
* Val   accuracy / confusion: 52.88% / [[30, 14, 2], [11, 21, 3], [7, 12, 4]]

------------------------------ Epoch 248 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.668556
- Iter 024 / 025, Loss: 0.875085

* Train accuracy / confusion: 64.00% / [[286, 55, 15], [102, 145, 20], [55, 41, 81]], 
* Val   accuracy / confusion: 46.15% / [[31, 12, 3], [16, 15, 4], [10, 11, 2]]

------------------------------ Epoch 249 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.584774
- Iter 024 / 025, Loss: 0.826705

* Train accuracy / confusion: 65.88% / [[297, 47, 12], [89, 149, 29], [44, 52, 81]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [10, 18, 7], [8, 11, 4]]

------------------------------ Epoch 250 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.760411
- Iter 024 / 025, Loss: 0.777268

* Train accuracy / confusion: 63.62% / [[283, 56, 14], [91, 152, 28], [54, 48, 74]], 
* Val   accuracy / confusion: 54.81% / [[30, 11, 5], [8, 22, 5], [9, 9, 5]]

------------------------------ Epoch 251 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.766131
- Iter 024 / 025, Loss: 0.727123

* Train accuracy / confusion: 65.25% / [[297, 48, 13], [86, 154, 29], [42, 60, 71]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [14, 18, 3], [5, 13, 5]]

------------------------------ Epoch 252 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.962595
- Iter 024 / 025, Loss: 0.831496

* Train accuracy / confusion: 66.38% / [[293, 56, 11], [90, 151, 26], [40, 46, 87]], 
* Val   accuracy / confusion: 52.88% / [[31, 10, 5], [12, 19, 4], [8, 10, 5]]

------------------------------ Epoch 253 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.939924
- Iter 024 / 025, Loss: 0.843104

* Train accuracy / confusion: 64.62% / [[288, 61, 13], [81, 151, 32], [45, 51, 78]], 
* Val   accuracy / confusion: 50.00% / [[30, 14, 2], [15, 13, 7], [5, 9, 9]]

------------------------------ Epoch 254 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.691719
- Iter 024 / 025, Loss: 0.720737

* Train accuracy / confusion: 66.50% / [[287, 53, 15], [78, 167, 22], [56, 44, 78]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [11, 20, 4], [6, 12, 5]]

------------------------------ Epoch 255 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.809415
- Iter 024 / 025, Loss: 0.790626

* Train accuracy / confusion: 66.62% / [[302, 44, 12], [85, 160, 20], [52, 54, 71]], 
* Val   accuracy / confusion: 57.69% / [[36, 9, 1], [10, 21, 4], [8, 12, 3]]

------------------------------ Epoch 256 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.704530
- Iter 024 / 025, Loss: 0.722232

* Train accuracy / confusion: 63.88% / [[297, 49, 13], [87, 150, 30], [49, 61, 64]], 
* Val   accuracy / confusion: 50.96% / [[30, 13, 3], [11, 19, 5], [7, 12, 4]]

------------------------------ Epoch 257 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.712040
- Iter 024 / 025, Loss: 0.820900

* Train accuracy / confusion: 65.62% / [[292, 46, 11], [88, 156, 29], [48, 53, 77]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [7, 23, 5], [5, 16, 2]]

------------------------------ Epoch 258 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.712087
- Iter 024 / 025, Loss: 0.827211

* Train accuracy / confusion: 67.00% / [[287, 60, 7], [71, 170, 26], [38, 62, 79]], 
* Val   accuracy / confusion: 50.96% / [[34, 10, 2], [13, 14, 8], [7, 11, 5]]

------------------------------ Epoch 259 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.711241
- Iter 024 / 025, Loss: 0.865462

* Train accuracy / confusion: 64.88% / [[291, 59, 9], [89, 156, 23], [49, 52, 72]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [9, 25, 1], [5, 16, 2]]

------------------------------ Epoch 260 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.778440
- Iter 024 / 025, Loss: 0.635812

* Train accuracy / confusion: 64.62% / [[289, 62, 7], [80, 152, 33], [45, 56, 76]], 
* Val   accuracy / confusion: 59.62% / [[34, 10, 2], [9, 21, 5], [10, 6, 7]]

------------------------------ Epoch 261 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.841277
- Iter 024 / 025, Loss: 0.781157

* Train accuracy / confusion: 63.88% / [[279, 57, 14], [94, 158, 21], [56, 47, 74]], 
* Val   accuracy / confusion: 57.69% / [[33, 11, 2], [8, 20, 7], [7, 9, 7]]

------------------------------ Epoch 262 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.735598
- Iter 024 / 025, Loss: 0.598780

* Train accuracy / confusion: 67.75% / [[307, 48, 5], [88, 156, 24], [42, 51, 79]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [12, 19, 4], [7, 10, 6]]

------------------------------ Epoch 263 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 1.090454
- Iter 024 / 025, Loss: 0.819181

* Train accuracy / confusion: 66.50% / [[297, 47, 11], [88, 160, 20], [40, 62, 75]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [11, 19, 5], [9, 10, 4]]

------------------------------ Epoch 264 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.672797
- Iter 024 / 025, Loss: 0.879051

* Train accuracy / confusion: 67.25% / [[301, 42, 12], [93, 153, 23], [47, 45, 84]], 
* Val   accuracy / confusion: 55.77% / [[36, 8, 2], [12, 16, 7], [6, 11, 6]]

------------------------------ Epoch 265 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.873235
- Iter 024 / 025, Loss: 1.176870

* Train accuracy / confusion: 61.00% / [[283, 61, 13], [91, 139, 34], [52, 61, 66]], 
* Val   accuracy / confusion: 53.85% / [[31, 13, 2], [10, 21, 4], [6, 13, 4]]

------------------------------ Epoch 266 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.610082
- Iter 024 / 025, Loss: 1.004577

* Train accuracy / confusion: 64.62% / [[285, 52, 19], [85, 154, 31], [52, 44, 78]], 
* Val   accuracy / confusion: 56.73% / [[34, 11, 1], [10, 20, 5], [7, 11, 5]]

------------------------------ Epoch 267 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.861344
- Iter 024 / 025, Loss: 0.904576

* Train accuracy / confusion: 66.25% / [[296, 46, 15], [84, 152, 35], [45, 45, 82]], 
* Val   accuracy / confusion: 54.81% / [[35, 8, 3], [12, 16, 7], [9, 8, 6]]

------------------------------ Epoch 268 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.639430
- Iter 024 / 025, Loss: 0.627686

* Train accuracy / confusion: 66.50% / [[299, 49, 8], [86, 157, 26], [49, 50, 76]], 
* Val   accuracy / confusion: 60.58% / [[34, 10, 2], [10, 21, 4], [9, 6, 8]]

------------------------------ Epoch 269 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.992834
- Iter 024 / 025, Loss: 0.930063

* Train accuracy / confusion: 64.25% / [[278, 63, 8], [85, 159, 28], [49, 53, 77]], 
* Val   accuracy / confusion: 56.73% / [[34, 12, 0], [11, 21, 3], [7, 12, 4]]

------------------------------ Epoch 270 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.660833
- Iter 024 / 025, Loss: 0.799033

* Train accuracy / confusion: 65.12% / [[293, 56, 7], [98, 148, 29], [44, 45, 80]], 
* Val   accuracy / confusion: 60.58% / [[33, 12, 1], [7, 25, 3], [7, 11, 5]]

------------------------------ Epoch 271 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.605187
- Iter 024 / 025, Loss: 0.700804

* Train accuracy / confusion: 66.50% / [[299, 51, 12], [86, 156, 24], [49, 46, 77]], 
* Val   accuracy / confusion: 49.04% / [[30, 15, 1], [11, 18, 6], [5, 15, 3]]

------------------------------ Epoch 272 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.826172
- Iter 024 / 025, Loss: 0.659994

* Train accuracy / confusion: 67.75% / [[298, 48, 11], [83, 161, 22], [42, 52, 83]], 
* Val   accuracy / confusion: 56.73% / [[33, 11, 2], [9, 19, 7], [6, 10, 7]]

------------------------------ Epoch 273 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.674074
- Iter 024 / 025, Loss: 0.657424

* Train accuracy / confusion: 64.25% / [[281, 59, 13], [83, 161, 26], [49, 56, 72]], 
* Val   accuracy / confusion: 61.54% / [[34, 10, 2], [8, 24, 3], [9, 8, 6]]

------------------------------ Epoch 274 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.621051
- Iter 024 / 025, Loss: 0.745407

* Train accuracy / confusion: 65.62% / [[294, 56, 8], [88, 154, 27], [38, 58, 77]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [10, 19, 6], [8, 10, 5]]

------------------------------ Epoch 275 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.848049
- Iter 024 / 025, Loss: 0.995819

* Train accuracy / confusion: 67.00% / [[301, 46, 11], [89, 154, 25], [49, 44, 81]], 
* Val   accuracy / confusion: 53.85% / [[35, 9, 2], [11, 18, 6], [6, 14, 3]]

------------------------------ Epoch 276 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.694524
- Iter 024 / 025, Loss: 0.749219

* Train accuracy / confusion: 65.38% / [[287, 61, 6], [83, 158, 26], [49, 52, 78]], 
* Val   accuracy / confusion: 52.88% / [[34, 10, 2], [10, 18, 7], [10, 10, 3]]

------------------------------ Epoch 277 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.686842
- Iter 024 / 025, Loss: 0.715204

* Train accuracy / confusion: 65.88% / [[296, 52, 11], [91, 151, 25], [43, 51, 80]], 
* Val   accuracy / confusion: 52.88% / [[35, 9, 2], [13, 14, 8], [4, 13, 6]]

------------------------------ Epoch 278 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.922502
- Iter 024 / 025, Loss: 0.661882

* Train accuracy / confusion: 64.50% / [[289, 57, 11], [88, 146, 33], [46, 49, 81]], 
* Val   accuracy / confusion: 56.73% / [[30, 14, 2], [8, 23, 4], [9, 8, 6]]

------------------------------ Epoch 279 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.840111
- Iter 024 / 025, Loss: 0.765245

* Train accuracy / confusion: 64.50% / [[289, 57, 12], [89, 145, 33], [40, 53, 82]], 
* Val   accuracy / confusion: 57.69% / [[33, 11, 2], [12, 20, 3], [9, 7, 7]]

------------------------------ Epoch 280 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.748833
- Iter 024 / 025, Loss: 0.707682

* Train accuracy / confusion: 63.62% / [[296, 49, 12], [102, 138, 28], [42, 58, 75]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [13, 17, 5], [7, 9, 7]]

------------------------------ Epoch 281 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.644881
- Iter 024 / 025, Loss: 0.774810

* Train accuracy / confusion: 65.75% / [[296, 47, 11], [87, 156, 27], [49, 53, 74]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [10, 20, 5], [7, 9, 7]]

------------------------------ Epoch 282 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.675104
- Iter 024 / 025, Loss: 0.797500

* Train accuracy / confusion: 66.12% / [[290, 56, 9], [85, 160, 22], [40, 59, 79]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [9, 23, 3], [9, 12, 2]]

------------------------------ Epoch 283 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.818131
- Iter 024 / 025, Loss: 0.574371

* Train accuracy / confusion: 66.88% / [[294, 52, 15], [81, 159, 27], [43, 47, 82]], 
* Val   accuracy / confusion: 55.77% / [[30, 14, 2], [12, 22, 1], [9, 8, 6]]

------------------------------ Epoch 284 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.839836
- Iter 024 / 025, Loss: 0.656550

* Train accuracy / confusion: 65.38% / [[284, 55, 13], [86, 159, 24], [49, 50, 80]], 
* Val   accuracy / confusion: 55.77% / [[34, 10, 2], [8, 18, 9], [3, 14, 6]]

------------------------------ Epoch 285 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.780016
- Iter 024 / 025, Loss: 0.770204

* Train accuracy / confusion: 66.38% / [[287, 53, 18], [79, 163, 25], [48, 46, 81]], 
* Val   accuracy / confusion: 57.69% / [[33, 11, 2], [10, 21, 4], [8, 9, 6]]

------------------------------ Epoch 286 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.821362
- Iter 024 / 025, Loss: 0.645732

* Train accuracy / confusion: 66.25% / [[288, 56, 13], [85, 157, 28], [43, 45, 85]], 
* Val   accuracy / confusion: 56.73% / [[33, 9, 4], [11, 18, 6], [8, 7, 8]]

------------------------------ Epoch 287 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.865343
- Iter 024 / 025, Loss: 0.862841

* Train accuracy / confusion: 66.88% / [[293, 53, 9], [88, 158, 25], [50, 40, 84]], 
* Val   accuracy / confusion: 60.58% / [[32, 14, 0], [8, 25, 2], [5, 12, 6]]

------------------------------ Epoch 288 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.870759
- Iter 024 / 025, Loss: 0.894996

* Train accuracy / confusion: 67.50% / [[296, 46, 10], [78, 168, 23], [38, 65, 76]], 
* Val   accuracy / confusion: 55.77% / [[32, 12, 2], [10, 23, 2], [10, 10, 3]]

------------------------------ Epoch 289 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.815072
- Iter 024 / 025, Loss: 0.772837

* Train accuracy / confusion: 66.50% / [[296, 52, 13], [81, 158, 27], [50, 45, 78]], 
* Val   accuracy / confusion: 58.65% / [[32, 12, 2], [11, 22, 2], [8, 8, 7]]

------------------------------ Epoch 290 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.799486
- Iter 024 / 025, Loss: 0.742707

* Train accuracy / confusion: 68.50% / [[297, 43, 11], [79, 165, 26], [50, 43, 86]], 
* Val   accuracy / confusion: 54.81% / [[37, 6, 3], [12, 16, 7], [4, 15, 4]]

------------------------------ Epoch 291 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.847889
- Iter 024 / 025, Loss: 0.659721

* Train accuracy / confusion: 63.75% / [[289, 50, 17], [94, 144, 31], [45, 53, 77]], 
* Val   accuracy / confusion: 54.81% / [[34, 8, 4], [13, 19, 3], [10, 9, 4]]

------------------------------ Epoch 292 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.755107
- Iter 024 / 025, Loss: 0.642016

* Train accuracy / confusion: 65.38% / [[289, 60, 9], [87, 156, 23], [43, 55, 78]], 
* Val   accuracy / confusion: 54.81% / [[32, 14, 0], [10, 20, 5], [6, 12, 5]]

------------------------------ Epoch 293 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.677451
- Iter 024 / 025, Loss: 0.543065

* Train accuracy / confusion: 68.25% / [[297, 44, 13], [81, 163, 25], [41, 50, 86]], 
* Val   accuracy / confusion: 51.92% / [[34, 12, 0], [10, 17, 8], [8, 12, 3]]

------------------------------ Epoch 294 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.634812
- Iter 024 / 025, Loss: 0.768019

* Train accuracy / confusion: 65.38% / [[280, 57, 12], [79, 167, 26], [43, 60, 76]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [12, 19, 4], [15, 5, 3]]

------------------------------ Epoch 295 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.906791
- Iter 024 / 025, Loss: 0.778981

* Train accuracy / confusion: 69.00% / [[297, 44, 14], [81, 167, 20], [42, 47, 88]], 
* Val   accuracy / confusion: 47.12% / [[32, 11, 3], [15, 15, 5], [9, 12, 2]]

------------------------------ Epoch 296 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.624379
- Iter 024 / 025, Loss: 0.713902

* Train accuracy / confusion: 68.00% / [[292, 54, 10], [82, 167, 20], [46, 44, 85]], 
* Val   accuracy / confusion: 50.96% / [[31, 13, 2], [12, 19, 4], [7, 13, 3]]

------------------------------ Epoch 297 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.685348
- Iter 024 / 025, Loss: 0.675968

* Train accuracy / confusion: 64.00% / [[283, 61, 14], [86, 151, 27], [51, 49, 78]], 
* Val   accuracy / confusion: 53.85% / [[34, 8, 4], [10, 17, 8], [9, 9, 5]]

------------------------------ Epoch 298 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.680174
- Iter 024 / 025, Loss: 0.627282

* Train accuracy / confusion: 66.50% / [[301, 42, 10], [94, 151, 27], [52, 43, 80]], 
* Val   accuracy / confusion: 52.88% / [[29, 13, 4], [9, 22, 4], [8, 11, 4]]

------------------------------ Epoch 299 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.580510
- Iter 024 / 025, Loss: 0.684175

* Train accuracy / confusion: 66.12% / [[285, 58, 11], [87, 158, 25], [45, 45, 86]], 
* Val   accuracy / confusion: 50.00% / [[31, 14, 1], [12, 19, 4], [9, 12, 2]]

------------------------------ Epoch 300 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 1.114863
- Iter 024 / 025, Loss: 0.711103

* Train accuracy / confusion: 66.50% / [[292, 51, 10], [72, 164, 34], [51, 50, 76]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [10, 17, 8], [9, 9, 5]]

------------------------------ Epoch 301 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.901580
- Iter 024 / 025, Loss: 0.845188

* Train accuracy / confusion: 66.12% / [[284, 62, 7], [88, 156, 26], [45, 43, 89]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [9, 18, 8], [9, 10, 4]]

------------------------------ Epoch 302 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.677212
- Iter 024 / 025, Loss: 0.673218

* Train accuracy / confusion: 66.00% / [[289, 52, 14], [78, 165, 26], [50, 52, 74]], 
* Val   accuracy / confusion: 54.81% / [[37, 8, 1], [13, 16, 6], [8, 11, 4]]

------------------------------ Epoch 303 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.826591
- Iter 024 / 025, Loss: 0.741152

* Train accuracy / confusion: 66.62% / [[292, 51, 10], [83, 169, 21], [49, 53, 72]], 
* Val   accuracy / confusion: 56.73% / [[31, 11, 4], [6, 25, 4], [6, 14, 3]]

------------------------------ Epoch 304 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.787878
- Iter 024 / 025, Loss: 0.900021

* Train accuracy / confusion: 65.75% / [[281, 58, 14], [87, 162, 22], [48, 45, 83]], 
* Val   accuracy / confusion: 51.92% / [[29, 15, 2], [13, 19, 3], [6, 11, 6]]

------------------------------ Epoch 305 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.698002
- Iter 024 / 025, Loss: 0.655440

* Train accuracy / confusion: 66.00% / [[288, 53, 14], [76, 164, 31], [43, 55, 76]], 
* Val   accuracy / confusion: 49.04% / [[29, 16, 1], [10, 17, 8], [5, 13, 5]]

------------------------------ Epoch 306 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.813954
- Iter 024 / 025, Loss: 0.735637

* Train accuracy / confusion: 67.38% / [[287, 53, 16], [88, 162, 16], [47, 41, 90]], 
* Val   accuracy / confusion: 51.92% / [[34, 10, 2], [11, 18, 6], [10, 11, 2]]

------------------------------ Epoch 307 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.785994
- Iter 024 / 025, Loss: 0.848298

* Train accuracy / confusion: 67.88% / [[293, 56, 12], [70, 165, 31], [39, 49, 85]], 
* Val   accuracy / confusion: 55.77% / [[36, 6, 4], [12, 16, 7], [9, 8, 6]]

------------------------------ Epoch 308 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.583412
- Iter 024 / 025, Loss: 0.608475

* Train accuracy / confusion: 68.62% / [[297, 42, 18], [79, 167, 24], [49, 39, 85]], 
* Val   accuracy / confusion: 53.85% / [[35, 9, 2], [10, 17, 8], [6, 13, 4]]

------------------------------ Epoch 309 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.916700
- Iter 024 / 025, Loss: 0.771069

* Train accuracy / confusion: 65.38% / [[289, 45, 21], [96, 143, 29], [42, 44, 91]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [9, 21, 5], [10, 9, 4]]

------------------------------ Epoch 310 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 1.031375
- Iter 024 / 025, Loss: 0.786993

* Train accuracy / confusion: 67.12% / [[289, 56, 13], [75, 163, 25], [47, 47, 85]], 
* Val   accuracy / confusion: 51.92% / [[31, 11, 4], [11, 20, 4], [9, 11, 3]]

------------------------------ Epoch 311 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.788166
- Iter 024 / 025, Loss: 0.756422

* Train accuracy / confusion: 66.75% / [[285, 56, 14], [84, 160, 28], [43, 41, 89]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [11, 18, 6], [7, 13, 3]]

------------------------------ Epoch 312 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.906305
- Iter 024 / 025, Loss: 0.733726

* Train accuracy / confusion: 67.12% / [[289, 52, 15], [82, 167, 21], [45, 48, 81]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [11, 22, 2], [6, 15, 2]]

------------------------------ Epoch 313 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.794048
- Iter 024 / 025, Loss: 0.701878

* Train accuracy / confusion: 66.88% / [[296, 48, 10], [83, 164, 22], [45, 57, 75]], 
* Val   accuracy / confusion: 47.12% / [[31, 11, 4], [11, 16, 8], [10, 11, 2]]

------------------------------ Epoch 314 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.617855
- Iter 024 / 025, Loss: 0.591887

* Train accuracy / confusion: 66.62% / [[295, 49, 14], [85, 156, 26], [48, 45, 82]], 
* Val   accuracy / confusion: 58.65% / [[35, 7, 4], [11, 19, 5], [7, 9, 7]]

------------------------------ Epoch 315 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.780531
- Iter 024 / 025, Loss: 0.681528

* Train accuracy / confusion: 66.38% / [[285, 57, 15], [86, 161, 22], [48, 41, 85]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [9, 21, 5], [8, 10, 5]]

------------------------------ Epoch 316 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.613416
- Iter 024 / 025, Loss: 0.677476

* Train accuracy / confusion: 67.50% / [[294, 49, 11], [75, 163, 30], [44, 51, 83]], 
* Val   accuracy / confusion: 52.88% / [[34, 9, 3], [11, 18, 6], [6, 14, 3]]

------------------------------ Epoch 317 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 1.048948
- Iter 024 / 025, Loss: 0.641996

* Train accuracy / confusion: 67.50% / [[301, 44, 12], [87, 157, 25], [40, 52, 82]], 
* Val   accuracy / confusion: 55.77% / [[31, 12, 3], [11, 18, 6], [6, 8, 9]]

------------------------------ Epoch 318 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.589272
- Iter 024 / 025, Loss: 0.738347

* Train accuracy / confusion: 69.00% / [[298, 46, 12], [76, 173, 19], [45, 50, 81]], 
* Val   accuracy / confusion: 49.04% / [[31, 14, 1], [10, 16, 9], [7, 12, 4]]

------------------------------ Epoch 319 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.597147
- Iter 024 / 025, Loss: 0.765219

* Train accuracy / confusion: 66.88% / [[294, 49, 11], [87, 160, 22], [45, 51, 81]], 
* Val   accuracy / confusion: 55.77% / [[31, 12, 3], [10, 21, 4], [7, 10, 6]]

------------------------------ Epoch 320 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.880515
- Iter 024 / 025, Loss: 0.591838

* Train accuracy / confusion: 67.38% / [[301, 48, 9], [82, 149, 36], [46, 40, 89]], 
* Val   accuracy / confusion: 53.85% / [[33, 13, 0], [12, 18, 5], [7, 11, 5]]

------------------------------ Epoch 321 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.930680
- Iter 024 / 025, Loss: 0.636665

* Train accuracy / confusion: 68.50% / [[300, 45, 6], [88, 164, 19], [47, 47, 84]], 
* Val   accuracy / confusion: 56.73% / [[32, 10, 4], [8, 20, 7], [7, 9, 7]]

------------------------------ Epoch 322 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.687057
- Iter 024 / 025, Loss: 0.914155

* Train accuracy / confusion: 67.62% / [[288, 55, 12], [77, 163, 29], [45, 41, 90]], 
* Val   accuracy / confusion: 57.69% / [[31, 14, 1], [12, 20, 3], [8, 6, 9]]

------------------------------ Epoch 323 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.747991
- Iter 024 / 025, Loss: 0.632263

* Train accuracy / confusion: 67.12% / [[290, 47, 17], [76, 168, 29], [44, 50, 79]], 
* Val   accuracy / confusion: 52.88% / [[33, 12, 1], [14, 18, 3], [10, 9, 4]]

------------------------------ Epoch 324 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.816032
- Iter 024 / 025, Loss: 0.935285

* Train accuracy / confusion: 70.25% / [[298, 51, 11], [68, 175, 24], [37, 47, 89]], 
* Val   accuracy / confusion: 55.77% / [[31, 14, 1], [8, 21, 6], [3, 14, 6]]

------------------------------ Epoch 325 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.799555
- Iter 024 / 025, Loss: 0.597779

* Train accuracy / confusion: 68.00% / [[298, 43, 13], [73, 168, 27], [51, 49, 78]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [11, 21, 3], [6, 12, 5]]

------------------------------ Epoch 326 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.771053
- Iter 024 / 025, Loss: 0.824393

* Train accuracy / confusion: 67.62% / [[286, 61, 5], [79, 168, 23], [49, 42, 87]], 
* Val   accuracy / confusion: 56.73% / [[34, 11, 1], [12, 20, 3], [4, 14, 5]]

------------------------------ Epoch 327 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.793472
- Iter 024 / 025, Loss: 0.807334

* Train accuracy / confusion: 67.50% / [[294, 47, 10], [88, 162, 22], [38, 55, 84]], 
* Val   accuracy / confusion: 54.81% / [[33, 12, 1], [11, 20, 4], [10, 9, 4]]

------------------------------ Epoch 328 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.813631
- Iter 024 / 025, Loss: 0.907786

* Train accuracy / confusion: 67.62% / [[292, 49, 13], [79, 166, 26], [44, 48, 83]], 
* Val   accuracy / confusion: 57.69% / [[33, 11, 2], [9, 21, 5], [7, 10, 6]]

------------------------------ Epoch 329 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.499038
- Iter 024 / 025, Loss: 0.752502

* Train accuracy / confusion: 68.38% / [[292, 54, 9], [87, 163, 19], [43, 41, 92]], 
* Val   accuracy / confusion: 53.85% / [[29, 13, 4], [10, 22, 3], [7, 11, 5]]

------------------------------ Epoch 330 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.623769
- Iter 024 / 025, Loss: 0.476061

* Train accuracy / confusion: 69.38% / [[299, 42, 19], [69, 166, 29], [37, 49, 90]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [9, 21, 5], [7, 12, 4]]

------------------------------ Epoch 331 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.754869
- Iter 024 / 025, Loss: 0.598784

* Train accuracy / confusion: 70.00% / [[300, 51, 5], [75, 164, 30], [42, 37, 96]], 
* Val   accuracy / confusion: 53.85% / [[34, 11, 1], [10, 18, 7], [3, 16, 4]]

------------------------------ Epoch 332 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.843727
- Iter 024 / 025, Loss: 0.622211

* Train accuracy / confusion: 66.50% / [[284, 56, 15], [85, 162, 19], [44, 49, 86]], 
* Val   accuracy / confusion: 58.65% / [[33, 10, 3], [11, 22, 2], [7, 10, 6]]

------------------------------ Epoch 333 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.736271
- Iter 024 / 025, Loss: 0.718841

* Train accuracy / confusion: 66.00% / [[290, 60, 14], [81, 154, 28], [38, 51, 84]], 
* Val   accuracy / confusion: 56.73% / [[31, 13, 2], [10, 23, 2], [6, 12, 5]]

------------------------------ Epoch 334 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.801185
- Iter 024 / 025, Loss: 0.723803

* Train accuracy / confusion: 68.38% / [[304, 44, 9], [87, 150, 30], [35, 48, 93]], 
* Val   accuracy / confusion: 50.96% / [[33, 11, 2], [15, 15, 5], [5, 13, 5]]

------------------------------ Epoch 335 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.874255
- Iter 024 / 025, Loss: 0.817342

* Train accuracy / confusion: 69.12% / [[298, 42, 13], [84, 162, 25], [40, 43, 93]], 
* Val   accuracy / confusion: 50.00% / [[30, 12, 4], [14, 15, 6], [4, 12, 7]]

------------------------------ Epoch 336 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.572263
- Iter 024 / 025, Loss: 0.660914

* Train accuracy / confusion: 68.00% / [[290, 56, 10], [81, 165, 24], [40, 45, 89]], 
* Val   accuracy / confusion: 50.00% / [[29, 13, 4], [9, 18, 8], [3, 15, 5]]

------------------------------ Epoch 337 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.490759
- Iter 024 / 025, Loss: 0.701358

* Train accuracy / confusion: 67.38% / [[289, 53, 13], [83, 167, 19], [45, 48, 83]], 
* Val   accuracy / confusion: 53.85% / [[30, 13, 3], [10, 21, 4], [5, 13, 5]]

------------------------------ Epoch 338 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.745914
- Iter 024 / 025, Loss: 0.772724

* Train accuracy / confusion: 67.50% / [[293, 53, 10], [80, 159, 28], [43, 46, 88]], 
* Val   accuracy / confusion: 48.08% / [[30, 11, 5], [13, 15, 7], [7, 11, 5]]

------------------------------ Epoch 339 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.675025
- Iter 024 / 025, Loss: 0.625643

* Train accuracy / confusion: 65.75% / [[291, 49, 14], [87, 150, 33], [45, 46, 85]], 
* Val   accuracy / confusion: 50.00% / [[28, 16, 2], [10, 20, 5], [4, 15, 4]]

------------------------------ Epoch 340 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.746775
- Iter 024 / 025, Loss: 0.926869

* Train accuracy / confusion: 67.62% / [[294, 47, 13], [82, 162, 23], [47, 47, 85]], 
* Val   accuracy / confusion: 56.73% / [[33, 12, 1], [12, 20, 3], [5, 12, 6]]

------------------------------ Epoch 341 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.618824
- Iter 024 / 025, Loss: 0.683344

* Train accuracy / confusion: 71.00% / [[302, 46, 9], [70, 176, 23], [42, 42, 90]], 
* Val   accuracy / confusion: 48.08% / [[30, 14, 2], [13, 18, 4], [4, 17, 2]]

------------------------------ Epoch 342 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.942432
- Iter 024 / 025, Loss: 0.799417

* Train accuracy / confusion: 68.50% / [[290, 55, 10], [80, 171, 18], [41, 48, 87]], 
* Val   accuracy / confusion: 56.73% / [[33, 13, 0], [12, 21, 2], [6, 12, 5]]

------------------------------ Epoch 343 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.653164
- Iter 024 / 025, Loss: 0.503158

* Train accuracy / confusion: 69.88% / [[308, 36, 12], [74, 172, 26], [45, 48, 79]], 
* Val   accuracy / confusion: 50.96% / [[33, 12, 1], [13, 16, 6], [6, 13, 4]]

------------------------------ Epoch 344 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.540160
- Iter 024 / 025, Loss: 0.851963

* Train accuracy / confusion: 69.25% / [[295, 51, 11], [78, 177, 16], [42, 48, 82]], 
* Val   accuracy / confusion: 48.08% / [[30, 14, 2], [12, 18, 5], [11, 10, 2]]

------------------------------ Epoch 345 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.630833
- Iter 024 / 025, Loss: 0.899169

* Train accuracy / confusion: 69.38% / [[303, 43, 9], [80, 162, 29], [37, 47, 90]], 
* Val   accuracy / confusion: 53.85% / [[32, 13, 1], [12, 18, 5], [8, 9, 6]]

------------------------------ Epoch 346 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.680159
- Iter 024 / 025, Loss: 0.686183

* Train accuracy / confusion: 68.75% / [[295, 50, 11], [81, 173, 19], [42, 47, 82]], 
* Val   accuracy / confusion: 54.81% / [[33, 12, 1], [10, 22, 3], [10, 11, 2]]

------------------------------ Epoch 347 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.718680
- Iter 024 / 025, Loss: 0.553908

* Train accuracy / confusion: 68.38% / [[302, 43, 14], [85, 162, 22], [46, 43, 83]], 
* Val   accuracy / confusion: 52.88% / [[33, 13, 0], [12, 18, 5], [10, 9, 4]]

------------------------------ Epoch 348 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.781727
- Iter 024 / 025, Loss: 0.759065

* Train accuracy / confusion: 69.50% / [[307, 44, 7], [71, 164, 31], [40, 51, 85]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [11, 21, 3], [6, 11, 6]]

------------------------------ Epoch 349 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.734960
- Iter 024 / 025, Loss: 0.683593

* Train accuracy / confusion: 67.88% / [[296, 49, 13], [87, 153, 25], [39, 44, 94]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [12, 17, 6], [6, 10, 7]]

------------------------------ Epoch 350 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.760463
- Iter 024 / 025, Loss: 0.662001

* Train accuracy / confusion: 66.75% / [[299, 41, 18], [93, 152, 23], [48, 43, 83]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [10, 21, 4], [8, 13, 2]]

------------------------------ Epoch 351 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.510073
- Iter 024 / 025, Loss: 0.729017

* Train accuracy / confusion: 68.38% / [[296, 44, 13], [86, 162, 23], [48, 39, 89]], 
* Val   accuracy / confusion: 54.81% / [[31, 11, 4], [8, 23, 4], [7, 13, 3]]

------------------------------ Epoch 352 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.691033
- Iter 024 / 025, Loss: 0.905965

* Train accuracy / confusion: 68.88% / [[295, 50, 11], [76, 173, 21], [37, 54, 83]], 
* Val   accuracy / confusion: 54.81% / [[33, 12, 1], [13, 17, 5], [4, 12, 7]]

------------------------------ Epoch 353 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.505527
- Iter 024 / 025, Loss: 0.852402

* Train accuracy / confusion: 71.00% / [[304, 40, 14], [67, 174, 27], [38, 46, 90]], 
* Val   accuracy / confusion: 52.88% / [[33, 8, 5], [9, 17, 9], [7, 11, 5]]

------------------------------ Epoch 354 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.690360
- Iter 024 / 025, Loss: 0.502778

* Train accuracy / confusion: 70.88% / [[300, 45, 10], [69, 176, 26], [37, 46, 91]], 
* Val   accuracy / confusion: 54.81% / [[31, 12, 3], [11, 20, 4], [10, 7, 6]]

------------------------------ Epoch 355 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.589140
- Iter 024 / 025, Loss: 0.465673

* Train accuracy / confusion: 68.62% / [[295, 45, 16], [82, 164, 21], [39, 48, 90]], 
* Val   accuracy / confusion: 54.81% / [[30, 12, 4], [11, 22, 2], [7, 11, 5]]

------------------------------ Epoch 356 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.772290
- Iter 024 / 025, Loss: 0.570642

* Train accuracy / confusion: 68.38% / [[292, 52, 13], [72, 167, 27], [43, 46, 88]], 
* Val   accuracy / confusion: 58.65% / [[34, 11, 1], [7, 22, 6], [7, 11, 5]]

------------------------------ Epoch 357 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.753118
- Iter 024 / 025, Loss: 0.651367

* Train accuracy / confusion: 68.25% / [[292, 47, 16], [82, 173, 18], [51, 40, 81]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [10, 21, 4], [8, 9, 6]]

------------------------------ Epoch 358 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.679708
- Iter 024 / 025, Loss: 0.751717

* Train accuracy / confusion: 68.25% / [[292, 53, 9], [78, 165, 27], [40, 47, 89]], 
* Val   accuracy / confusion: 60.58% / [[34, 9, 3], [8, 23, 4], [5, 12, 6]]

------------------------------ Epoch 359 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.688436
- Iter 024 / 025, Loss: 0.740065

* Train accuracy / confusion: 68.50% / [[297, 49, 10], [74, 170, 25], [42, 52, 81]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [9, 21, 5], [5, 15, 3]]

------------------------------ Epoch 360 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.677188
- Iter 024 / 025, Loss: 0.799582

* Train accuracy / confusion: 70.50% / [[301, 37, 19], [72, 172, 22], [44, 42, 91]], 
* Val   accuracy / confusion: 57.69% / [[32, 13, 1], [12, 20, 3], [6, 9, 8]]

------------------------------ Epoch 361 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.756229
- Iter 024 / 025, Loss: 0.612792

* Train accuracy / confusion: 68.88% / [[293, 48, 12], [74, 166, 32], [39, 44, 92]], 
* Val   accuracy / confusion: 60.58% / [[30, 14, 2], [6, 24, 5], [5, 9, 9]]

------------------------------ Epoch 362 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.596115
- Iter 024 / 025, Loss: 0.557012

* Train accuracy / confusion: 69.12% / [[288, 49, 17], [83, 172, 15], [33, 50, 93]], 
* Val   accuracy / confusion: 55.77% / [[31, 13, 2], [11, 22, 2], [8, 10, 5]]

------------------------------ Epoch 363 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.661822
- Iter 024 / 025, Loss: 0.605125

* Train accuracy / confusion: 70.00% / [[297, 46, 12], [75, 172, 19], [41, 47, 91]], 
* Val   accuracy / confusion: 55.77% / [[27, 16, 3], [6, 26, 3], [6, 12, 5]]

------------------------------ Epoch 364 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.724401
- Iter 024 / 025, Loss: 0.614671

* Train accuracy / confusion: 71.00% / [[299, 53, 8], [77, 171, 19], [35, 40, 98]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [9, 19, 7], [5, 11, 7]]

------------------------------ Epoch 365 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.659172
- Iter 024 / 025, Loss: 0.673774

* Train accuracy / confusion: 68.62% / [[303, 41, 11], [85, 161, 20], [41, 53, 85]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [10, 19, 6], [6, 12, 5]]

------------------------------ Epoch 366 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.593039
- Iter 024 / 025, Loss: 0.957503

* Train accuracy / confusion: 69.12% / [[297, 51, 10], [74, 169, 26], [43, 43, 87]], 
* Val   accuracy / confusion: 50.96% / [[30, 11, 5], [7, 19, 9], [6, 13, 4]]

------------------------------ Epoch 367 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.670691
- Iter 024 / 025, Loss: 0.754366

* Train accuracy / confusion: 71.50% / [[305, 41, 11], [68, 181, 18], [30, 60, 86]], 
* Val   accuracy / confusion: 55.77% / [[32, 12, 2], [7, 23, 5], [7, 13, 3]]

------------------------------ Epoch 368 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.716063
- Iter 024 / 025, Loss: 0.800534

* Train accuracy / confusion: 68.12% / [[300, 46, 6], [95, 153, 22], [38, 48, 92]], 
* Val   accuracy / confusion: 60.58% / [[34, 9, 3], [9, 23, 3], [6, 11, 6]]

------------------------------ Epoch 369 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.788079
- Iter 024 / 025, Loss: 0.649437

* Train accuracy / confusion: 66.38% / [[275, 64, 15], [79, 172, 20], [43, 48, 84]], 
* Val   accuracy / confusion: 55.77% / [[29, 14, 3], [7, 23, 5], [4, 13, 6]]

------------------------------ Epoch 370 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.791758
- Iter 024 / 025, Loss: 0.585909

* Train accuracy / confusion: 69.75% / [[299, 47, 11], [72, 170, 29], [38, 45, 89]], 
* Val   accuracy / confusion: 53.85% / [[26, 20, 0], [8, 22, 5], [6, 9, 8]]

------------------------------ Epoch 371 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.721903
- Iter 024 / 025, Loss: 0.759140

* Train accuracy / confusion: 70.00% / [[306, 38, 16], [79, 166, 22], [41, 44, 88]], 
* Val   accuracy / confusion: 53.85% / [[32, 14, 0], [9, 21, 5], [8, 12, 3]]

------------------------------ Epoch 372 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.540839
- Iter 024 / 025, Loss: 0.527425

* Train accuracy / confusion: 68.75% / [[292, 47, 16], [77, 170, 25], [35, 50, 88]], 
* Val   accuracy / confusion: 52.88% / [[31, 13, 2], [6, 21, 8], [9, 11, 3]]

------------------------------ Epoch 373 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.588413
- Iter 024 / 025, Loss: 0.632086

* Train accuracy / confusion: 68.62% / [[298, 51, 3], [81, 168, 22], [43, 51, 83]], 
* Val   accuracy / confusion: 54.81% / [[31, 15, 0], [9, 23, 3], [7, 13, 3]]

------------------------------ Epoch 374 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.655660
- Iter 024 / 025, Loss: 0.615892

* Train accuracy / confusion: 70.12% / [[298, 49, 7], [78, 175, 15], [36, 54, 88]], 
* Val   accuracy / confusion: 60.58% / [[32, 12, 2], [6, 24, 5], [5, 11, 7]]

------------------------------ Epoch 375 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.774186
- Iter 024 / 025, Loss: 1.060906

* Train accuracy / confusion: 69.00% / [[295, 54, 12], [79, 163, 25], [44, 34, 94]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [10, 20, 5], [8, 8, 7]]

------------------------------ Epoch 376 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.720283
- Iter 024 / 025, Loss: 0.690488

* Train accuracy / confusion: 73.12% / [[315, 37, 10], [72, 172, 21], [30, 45, 98]], 
* Val   accuracy / confusion: 53.85% / [[34, 11, 1], [10, 17, 8], [7, 11, 5]]

------------------------------ Epoch 377 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.579508
- Iter 024 / 025, Loss: 0.934797

* Train accuracy / confusion: 69.25% / [[298, 49, 10], [75, 167, 25], [37, 50, 89]], 
* Val   accuracy / confusion: 55.77% / [[36, 7, 3], [10, 18, 7], [6, 13, 4]]

------------------------------ Epoch 378 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.943933
- Iter 024 / 025, Loss: 0.505059

* Train accuracy / confusion: 69.75% / [[297, 46, 12], [78, 168, 19], [45, 42, 93]], 
* Val   accuracy / confusion: 53.85% / [[29, 17, 0], [13, 21, 1], [6, 11, 6]]

------------------------------ Epoch 379 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.752877
- Iter 024 / 025, Loss: 0.671175

* Train accuracy / confusion: 69.25% / [[289, 54, 12], [75, 167, 27], [41, 37, 98]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [9, 23, 3], [7, 12, 4]]

------------------------------ Epoch 380 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.566827
- Iter 024 / 025, Loss: 0.541163

* Train accuracy / confusion: 69.62% / [[297, 48, 16], [71, 167, 27], [28, 53, 93]], 
* Val   accuracy / confusion: 55.77% / [[31, 13, 2], [7, 22, 6], [6, 12, 5]]

------------------------------ Epoch 381 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.495707
- Iter 024 / 025, Loss: 0.700773

* Train accuracy / confusion: 70.50% / [[310, 39, 9], [77, 165, 24], [39, 48, 89]], 
* Val   accuracy / confusion: 63.46% / [[31, 14, 1], [5, 26, 4], [6, 8, 9]]

------------------------------ Epoch 382 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.713566
- Iter 024 / 025, Loss: 0.548567

* Train accuracy / confusion: 69.00% / [[300, 50, 7], [81, 159, 26], [33, 51, 93]], 
* Val   accuracy / confusion: 57.69% / [[33, 12, 1], [9, 22, 4], [7, 11, 5]]

------------------------------ Epoch 383 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.583214
- Iter 024 / 025, Loss: 0.775163

* Train accuracy / confusion: 69.50% / [[294, 46, 11], [77, 174, 18], [46, 46, 88]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [9, 20, 6], [7, 12, 4]]

------------------------------ Epoch 384 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.543807
- Iter 024 / 025, Loss: 0.706586

* Train accuracy / confusion: 72.88% / [[313, 39, 9], [68, 177, 19], [40, 42, 93]], 
* Val   accuracy / confusion: 52.88% / [[30, 13, 3], [11, 20, 4], [9, 9, 5]]

------------------------------ Epoch 385 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.646205
- Iter 024 / 025, Loss: 0.686560

* Train accuracy / confusion: 67.25% / [[286, 47, 20], [78, 167, 24], [42, 51, 85]], 
* Val   accuracy / confusion: 58.65% / [[32, 13, 1], [6, 25, 4], [8, 11, 4]]

------------------------------ Epoch 386 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.769939
- Iter 024 / 025, Loss: 0.577149

* Train accuracy / confusion: 71.88% / [[311, 39, 10], [71, 170, 24], [35, 46, 94]], 
* Val   accuracy / confusion: 53.85% / [[31, 13, 2], [9, 19, 7], [4, 13, 6]]

------------------------------ Epoch 387 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.666456
- Iter 024 / 025, Loss: 0.677832

* Train accuracy / confusion: 68.50% / [[287, 55, 13], [86, 163, 20], [33, 45, 98]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [7, 19, 9], [3, 13, 7]]

------------------------------ Epoch 388 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.774972
- Iter 024 / 025, Loss: 0.624763

* Train accuracy / confusion: 67.75% / [[295, 44, 13], [77, 161, 35], [39, 50, 86]], 
* Val   accuracy / confusion: 56.73% / [[33, 11, 2], [11, 20, 4], [3, 14, 6]]

------------------------------ Epoch 389 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.707621
- Iter 024 / 025, Loss: 0.747964

* Train accuracy / confusion: 71.12% / [[301, 47, 11], [71, 170, 24], [36, 42, 98]], 
* Val   accuracy / confusion: 56.73% / [[33, 12, 1], [11, 21, 3], [7, 11, 5]]

------------------------------ Epoch 390 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.649617
- Iter 024 / 025, Loss: 0.662061

* Train accuracy / confusion: 70.38% / [[305, 40, 10], [74, 172, 24], [38, 51, 86]], 
* Val   accuracy / confusion: 50.96% / [[32, 13, 1], [11, 17, 7], [7, 12, 4]]

------------------------------ Epoch 391 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.652160
- Iter 024 / 025, Loss: 0.882422

* Train accuracy / confusion: 66.88% / [[282, 57, 16], [87, 166, 18], [46, 41, 87]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [10, 19, 6], [6, 11, 6]]

------------------------------ Epoch 392 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.627003
- Iter 024 / 025, Loss: 0.694533

* Train accuracy / confusion: 71.00% / [[297, 44, 13], [66, 182, 25], [36, 48, 89]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [10, 20, 5], [8, 9, 6]]

------------------------------ Epoch 393 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.596864
- Iter 024 / 025, Loss: 0.567051

* Train accuracy / confusion: 70.00% / [[297, 43, 13], [76, 177, 19], [40, 49, 86]], 
* Val   accuracy / confusion: 56.73% / [[33, 10, 3], [11, 19, 5], [7, 9, 7]]

------------------------------ Epoch 394 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.668336
- Iter 024 / 025, Loss: 0.533805

* Train accuracy / confusion: 70.00% / [[284, 59, 9], [68, 178, 23], [43, 38, 98]], 
* Val   accuracy / confusion: 57.69% / [[30, 13, 3], [6, 25, 4], [5, 13, 5]]

------------------------------ Epoch 395 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.739501
- Iter 024 / 025, Loss: 0.526612

* Train accuracy / confusion: 68.50% / [[287, 52, 13], [82, 161, 30], [33, 42, 100]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [11, 19, 5], [6, 10, 7]]

------------------------------ Epoch 396 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.660202
- Iter 024 / 025, Loss: 0.602583

* Train accuracy / confusion: 71.38% / [[300, 42, 12], [72, 177, 18], [39, 46, 94]], 
* Val   accuracy / confusion: 52.88% / [[30, 13, 3], [11, 18, 6], [5, 11, 7]]

------------------------------ Epoch 397 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.652142
- Iter 024 / 025, Loss: 0.631560

* Train accuracy / confusion: 71.25% / [[288, 51, 19], [64, 192, 12], [41, 43, 90]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [11, 19, 5], [5, 11, 7]]

------------------------------ Epoch 398 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.477172
- Iter 024 / 025, Loss: 0.857012

* Train accuracy / confusion: 72.25% / [[308, 39, 12], [62, 182, 22], [40, 47, 88]], 
* Val   accuracy / confusion: 54.81% / [[32, 14, 0], [12, 20, 3], [6, 12, 5]]

------------------------------ Epoch 399 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.479858
- Iter 024 / 025, Loss: 0.836883

* Train accuracy / confusion: 70.88% / [[295, 44, 13], [76, 167, 23], [35, 42, 105]], 
* Val   accuracy / confusion: 50.96% / [[27, 16, 3], [10, 19, 6], [5, 11, 7]]

------------------------------ Epoch 400 / 500, Learning rate: 1.40e-03 ------------------------------
- Iter 012 / 025, Loss: 0.602104
- Iter 024 / 025, Loss: 0.466169

* Train accuracy / confusion: 70.25% / [[305, 43, 11], [79, 162, 25], [34, 46, 95]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [13, 20, 2], [8, 10, 5]]

------------------------------ Epoch 401 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.615853
- Iter 024 / 025, Loss: 0.495367

* Train accuracy / confusion: 70.12% / [[307, 36, 12], [88, 159, 22], [37, 44, 95]], 
* Val   accuracy / confusion: 59.62% / [[32, 10, 4], [9, 20, 6], [5, 8, 10]]

------------------------------ Epoch 402 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.568786
- Iter 024 / 025, Loss: 0.773237

* Train accuracy / confusion: 70.62% / [[300, 48, 11], [76, 172, 22], [38, 40, 93]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [9, 20, 6], [8, 8, 7]]

------------------------------ Epoch 403 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.655530
- Iter 024 / 025, Loss: 0.561953

* Train accuracy / confusion: 70.25% / [[313, 35, 10], [76, 165, 30], [42, 45, 84]], 
* Val   accuracy / confusion: 57.69% / [[34, 9, 3], [10, 19, 6], [9, 7, 7]]

------------------------------ Epoch 404 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.808159
- Iter 024 / 025, Loss: 0.603588

* Train accuracy / confusion: 70.38% / [[300, 43, 14], [82, 162, 23], [44, 31, 101]], 
* Val   accuracy / confusion: 61.54% / [[35, 9, 2], [9, 21, 5], [8, 7, 8]]

------------------------------ Epoch 405 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.525254
- Iter 024 / 025, Loss: 0.428517

* Train accuracy / confusion: 69.50% / [[300, 43, 14], [82, 161, 24], [41, 40, 95]], 
* Val   accuracy / confusion: 57.69% / [[31, 12, 3], [12, 20, 3], [4, 10, 9]]

------------------------------ Epoch 406 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.739927
- Iter 024 / 025, Loss: 0.684220

* Train accuracy / confusion: 72.25% / [[305, 38, 14], [74, 177, 14], [31, 51, 96]], 
* Val   accuracy / confusion: 51.92% / [[31, 11, 4], [13, 19, 3], [5, 14, 4]]

------------------------------ Epoch 407 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.573110
- Iter 024 / 025, Loss: 0.518988

* Train accuracy / confusion: 70.12% / [[300, 42, 15], [80, 165, 23], [35, 44, 96]], 
* Val   accuracy / confusion: 51.92% / [[33, 12, 1], [16, 16, 3], [10, 8, 5]]

------------------------------ Epoch 408 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.592063
- Iter 024 / 025, Loss: 0.605952

* Train accuracy / confusion: 70.38% / [[304, 47, 8], [81, 160, 28], [35, 38, 99]], 
* Val   accuracy / confusion: 53.85% / [[34, 10, 2], [13, 14, 8], [5, 10, 8]]

------------------------------ Epoch 409 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.596691
- Iter 024 / 025, Loss: 0.713742

* Train accuracy / confusion: 71.62% / [[301, 38, 14], [77, 170, 22], [34, 42, 102]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [13, 19, 3], [9, 9, 5]]

------------------------------ Epoch 410 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.560717
- Iter 024 / 025, Loss: 0.416063

* Train accuracy / confusion: 71.88% / [[309, 41, 12], [69, 171, 27], [43, 33, 95]], 
* Val   accuracy / confusion: 48.08% / [[30, 14, 2], [12, 18, 5], [5, 16, 2]]

------------------------------ Epoch 411 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.539378
- Iter 024 / 025, Loss: 0.789587

* Train accuracy / confusion: 70.75% / [[311, 41, 10], [76, 164, 25], [38, 44, 91]], 
* Val   accuracy / confusion: 57.69% / [[32, 12, 2], [10, 22, 3], [9, 8, 6]]

------------------------------ Epoch 412 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.601802
- Iter 024 / 025, Loss: 0.683195

* Train accuracy / confusion: 70.88% / [[300, 45, 9], [80, 163, 28], [27, 44, 104]], 
* Val   accuracy / confusion: 51.92% / [[30, 13, 3], [14, 16, 5], [4, 11, 8]]

------------------------------ Epoch 413 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.881243
- Iter 024 / 025, Loss: 0.851557

* Train accuracy / confusion: 71.00% / [[301, 46, 10], [67, 175, 25], [44, 40, 92]], 
* Val   accuracy / confusion: 54.81% / [[29, 17, 0], [10, 21, 4], [9, 7, 7]]

------------------------------ Epoch 414 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.912311
- Iter 024 / 025, Loss: 0.682042

* Train accuracy / confusion: 70.25% / [[300, 41, 12], [77, 165, 27], [38, 43, 97]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [11, 19, 5], [4, 12, 7]]

------------------------------ Epoch 415 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.583628
- Iter 024 / 025, Loss: 0.610365

* Train accuracy / confusion: 69.88% / [[298, 47, 9], [71, 170, 29], [41, 44, 91]], 
* Val   accuracy / confusion: 52.88% / [[29, 11, 6], [8, 20, 7], [8, 9, 6]]

------------------------------ Epoch 416 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.632298
- Iter 024 / 025, Loss: 0.588871

* Train accuracy / confusion: 72.00% / [[301, 45, 11], [71, 179, 21], [28, 48, 96]], 
* Val   accuracy / confusion: 49.04% / [[32, 13, 1], [13, 15, 7], [7, 12, 4]]

------------------------------ Epoch 417 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.524723
- Iter 024 / 025, Loss: 0.519227

* Train accuracy / confusion: 70.38% / [[301, 45, 14], [73, 169, 26], [33, 46, 93]], 
* Val   accuracy / confusion: 46.15% / [[28, 16, 2], [15, 16, 4], [7, 12, 4]]

------------------------------ Epoch 418 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.547789
- Iter 024 / 025, Loss: 0.710243

* Train accuracy / confusion: 70.12% / [[304, 44, 12], [76, 166, 25], [37, 45, 91]], 
* Val   accuracy / confusion: 50.00% / [[32, 12, 2], [13, 15, 7], [6, 12, 5]]

------------------------------ Epoch 419 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.575565
- Iter 024 / 025, Loss: 0.681588

* Train accuracy / confusion: 70.75% / [[307, 39, 9], [85, 163, 20], [38, 43, 96]], 
* Val   accuracy / confusion: 50.96% / [[32, 13, 1], [13, 19, 3], [6, 15, 2]]

------------------------------ Epoch 420 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.745658
- Iter 024 / 025, Loss: 0.615891

* Train accuracy / confusion: 71.00% / [[305, 44, 11], [69, 166, 30], [30, 48, 97]], 
* Val   accuracy / confusion: 58.65% / [[34, 8, 4], [14, 18, 3], [8, 6, 9]]

------------------------------ Epoch 421 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.711866
- Iter 024 / 025, Loss: 0.513466

* Train accuracy / confusion: 70.25% / [[301, 42, 9], [81, 168, 24], [35, 47, 93]], 
* Val   accuracy / confusion: 51.92% / [[31, 12, 3], [11, 19, 5], [10, 9, 4]]

------------------------------ Epoch 422 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.632982
- Iter 024 / 025, Loss: 0.662497

* Train accuracy / confusion: 69.00% / [[292, 46, 16], [76, 171, 22], [40, 48, 89]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [11, 19, 5], [5, 12, 6]]

------------------------------ Epoch 423 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.677314
- Iter 024 / 025, Loss: 0.663997

* Train accuracy / confusion: 71.00% / [[300, 42, 14], [73, 172, 23], [32, 48, 96]], 
* Val   accuracy / confusion: 48.08% / [[30, 14, 2], [14, 16, 5], [7, 12, 4]]

------------------------------ Epoch 424 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.900117
- Iter 024 / 025, Loss: 0.631496

* Train accuracy / confusion: 70.25% / [[305, 42, 6], [84, 158, 25], [41, 40, 99]], 
* Val   accuracy / confusion: 50.00% / [[28, 15, 3], [11, 17, 7], [6, 10, 7]]

------------------------------ Epoch 425 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.640252
- Iter 024 / 025, Loss: 0.617826

* Train accuracy / confusion: 72.50% / [[305, 40, 12], [73, 179, 18], [41, 36, 96]], 
* Val   accuracy / confusion: 56.73% / [[33, 10, 3], [11, 22, 2], [8, 11, 4]]

------------------------------ Epoch 426 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.548540
- Iter 024 / 025, Loss: 0.484508

* Train accuracy / confusion: 71.75% / [[308, 43, 6], [73, 169, 26], [34, 44, 97]], 
* Val   accuracy / confusion: 55.77% / [[31, 12, 3], [9, 21, 5], [7, 10, 6]]

------------------------------ Epoch 427 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.843293
- Iter 024 / 025, Loss: 0.806561

* Train accuracy / confusion: 68.75% / [[289, 45, 19], [72, 173, 26], [38, 50, 88]], 
* Val   accuracy / confusion: 53.85% / [[29, 14, 3], [8, 22, 5], [8, 10, 5]]

------------------------------ Epoch 428 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.699419
- Iter 024 / 025, Loss: 0.629469

* Train accuracy / confusion: 72.38% / [[302, 42, 10], [70, 178, 22], [35, 42, 99]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [9, 20, 6], [4, 14, 5]]

------------------------------ Epoch 429 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.817285
- Iter 024 / 025, Loss: 0.600721

* Train accuracy / confusion: 68.88% / [[291, 48, 17], [76, 172, 23], [37, 48, 88]], 
* Val   accuracy / confusion: 50.00% / [[32, 11, 3], [14, 16, 5], [7, 12, 4]]

------------------------------ Epoch 430 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.676629
- Iter 024 / 025, Loss: 0.659087

* Train accuracy / confusion: 72.12% / [[298, 44, 10], [75, 179, 20], [37, 37, 100]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [13, 15, 7], [7, 11, 5]]

------------------------------ Epoch 431 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.487588
- Iter 024 / 025, Loss: 0.802335

* Train accuracy / confusion: 70.00% / [[297, 45, 14], [76, 168, 24], [36, 45, 95]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [8, 22, 5], [7, 11, 5]]

------------------------------ Epoch 432 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.689796
- Iter 024 / 025, Loss: 0.741094

* Train accuracy / confusion: 73.38% / [[306, 43, 7], [64, 187, 15], [38, 46, 94]], 
* Val   accuracy / confusion: 50.96% / [[34, 10, 2], [15, 13, 7], [4, 13, 6]]

------------------------------ Epoch 433 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.676593
- Iter 024 / 025, Loss: 0.766858

* Train accuracy / confusion: 70.62% / [[305, 44, 10], [73, 169, 24], [43, 41, 91]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [10, 20, 5], [10, 10, 3]]

------------------------------ Epoch 434 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.763733
- Iter 024 / 025, Loss: 0.732788

* Train accuracy / confusion: 69.00% / [[293, 51, 10], [74, 167, 26], [42, 45, 92]], 
* Val   accuracy / confusion: 52.88% / [[31, 10, 5], [11, 20, 4], [5, 14, 4]]

------------------------------ Epoch 435 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.605941
- Iter 024 / 025, Loss: 0.660183

* Train accuracy / confusion: 72.88% / [[306, 36, 12], [72, 176, 19], [39, 39, 101]], 
* Val   accuracy / confusion: 57.69% / [[34, 10, 2], [10, 18, 7], [7, 8, 8]]

------------------------------ Epoch 436 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.577180
- Iter 024 / 025, Loss: 0.934311

* Train accuracy / confusion: 69.50% / [[303, 47, 11], [77, 160, 31], [34, 44, 93]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [11, 16, 8], [4, 13, 6]]

------------------------------ Epoch 437 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.695782
- Iter 024 / 025, Loss: 0.892589

* Train accuracy / confusion: 70.50% / [[303, 40, 10], [71, 179, 22], [40, 53, 82]], 
* Val   accuracy / confusion: 56.73% / [[34, 11, 1], [11, 20, 4], [7, 11, 5]]

------------------------------ Epoch 438 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.744356
- Iter 024 / 025, Loss: 0.636015

* Train accuracy / confusion: 69.25% / [[300, 48, 8], [76, 170, 24], [38, 52, 84]], 
* Val   accuracy / confusion: 50.96% / [[31, 13, 2], [13, 18, 4], [6, 13, 4]]

------------------------------ Epoch 439 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.465372
- Iter 024 / 025, Loss: 0.773002

* Train accuracy / confusion: 72.25% / [[308, 37, 8], [72, 177, 24], [49, 32, 93]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [9, 22, 4], [8, 11, 4]]

------------------------------ Epoch 440 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.950058
- Iter 024 / 025, Loss: 0.608921

* Train accuracy / confusion: 72.88% / [[312, 30, 13], [67, 179, 21], [37, 49, 92]], 
* Val   accuracy / confusion: 58.65% / [[32, 12, 2], [9, 21, 5], [6, 9, 8]]

------------------------------ Epoch 441 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.690704
- Iter 024 / 025, Loss: 0.736120

* Train accuracy / confusion: 69.62% / [[300, 39, 15], [73, 170, 29], [40, 47, 87]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [13, 15, 7], [6, 10, 7]]

------------------------------ Epoch 442 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.514224
- Iter 024 / 025, Loss: 0.768901

* Train accuracy / confusion: 69.88% / [[302, 44, 12], [71, 169, 24], [35, 55, 88]], 
* Val   accuracy / confusion: 53.85% / [[30, 13, 3], [11, 17, 7], [4, 10, 9]]

------------------------------ Epoch 443 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.471584
- Iter 024 / 025, Loss: 0.704810

* Train accuracy / confusion: 70.62% / [[301, 44, 14], [76, 166, 20], [33, 48, 98]], 
* Val   accuracy / confusion: 54.81% / [[34, 11, 1], [13, 20, 2], [12, 8, 3]]

------------------------------ Epoch 444 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.482860
- Iter 024 / 025, Loss: 0.608613

* Train accuracy / confusion: 71.88% / [[301, 45, 14], [81, 163, 19], [30, 36, 111]], 
* Val   accuracy / confusion: 54.81% / [[28, 13, 5], [10, 22, 3], [7, 9, 7]]

------------------------------ Epoch 445 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.631179
- Iter 024 / 025, Loss: 0.499534

* Train accuracy / confusion: 72.38% / [[305, 39, 11], [66, 180, 21], [38, 46, 94]], 
* Val   accuracy / confusion: 57.69% / [[30, 13, 3], [10, 23, 2], [10, 6, 7]]

------------------------------ Epoch 446 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.810181
- Iter 024 / 025, Loss: 0.681724

* Train accuracy / confusion: 70.62% / [[298, 51, 9], [73, 177, 19], [35, 48, 90]], 
* Val   accuracy / confusion: 52.88% / [[34, 11, 1], [14, 17, 4], [8, 11, 4]]

------------------------------ Epoch 447 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.538495
- Iter 024 / 025, Loss: 0.604844

* Train accuracy / confusion: 70.25% / [[296, 53, 6], [69, 174, 27], [30, 53, 92]], 
* Val   accuracy / confusion: 50.00% / [[32, 13, 1], [12, 15, 8], [7, 11, 5]]

------------------------------ Epoch 448 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.710753
- Iter 024 / 025, Loss: 0.553061

* Train accuracy / confusion: 70.50% / [[308, 43, 6], [80, 161, 27], [29, 51, 95]], 
* Val   accuracy / confusion: 59.62% / [[37, 7, 2], [9, 20, 6], [8, 10, 5]]

------------------------------ Epoch 449 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.570023
- Iter 024 / 025, Loss: 0.931810

* Train accuracy / confusion: 69.25% / [[294, 51, 10], [73, 172, 22], [48, 42, 88]], 
* Val   accuracy / confusion: 56.73% / [[33, 8, 5], [11, 21, 3], [7, 11, 5]]

------------------------------ Epoch 450 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.544663
- Iter 024 / 025, Loss: 0.589529

* Train accuracy / confusion: 70.62% / [[301, 40, 13], [78, 167, 26], [37, 41, 97]], 
* Val   accuracy / confusion: 56.73% / [[33, 10, 3], [8, 22, 5], [7, 12, 4]]

------------------------------ Epoch 451 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.802341
- Iter 024 / 025, Loss: 0.607151

* Train accuracy / confusion: 70.62% / [[293, 51, 14], [75, 177, 16], [35, 44, 95]], 
* Val   accuracy / confusion: 54.81% / [[31, 11, 4], [11, 20, 4], [6, 11, 6]]

------------------------------ Epoch 452 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.607343
- Iter 024 / 025, Loss: 1.179255

* Train accuracy / confusion: 70.75% / [[298, 46, 14], [76, 170, 20], [37, 41, 98]], 
* Val   accuracy / confusion: 58.65% / [[32, 11, 3], [11, 21, 3], [6, 9, 8]]

------------------------------ Epoch 453 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.514473
- Iter 024 / 025, Loss: 0.507122

* Train accuracy / confusion: 73.25% / [[310, 37, 10], [72, 178, 23], [30, 42, 98]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [12, 18, 5], [5, 14, 4]]

------------------------------ Epoch 454 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.781116
- Iter 024 / 025, Loss: 0.462467

* Train accuracy / confusion: 71.00% / [[300, 47, 11], [75, 175, 16], [34, 49, 93]], 
* Val   accuracy / confusion: 52.88% / [[30, 13, 3], [10, 20, 5], [9, 9, 5]]

------------------------------ Epoch 455 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.660803
- Iter 024 / 025, Loss: 0.614599

* Train accuracy / confusion: 70.38% / [[302, 46, 10], [74, 167, 26], [38, 43, 94]], 
* Val   accuracy / confusion: 54.81% / [[31, 11, 4], [11, 19, 5], [7, 9, 7]]

------------------------------ Epoch 456 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.428998
- Iter 024 / 025, Loss: 0.627170

* Train accuracy / confusion: 71.50% / [[295, 42, 18], [75, 182, 15], [36, 42, 95]], 
* Val   accuracy / confusion: 53.85% / [[36, 8, 2], [14, 16, 5], [4, 15, 4]]

------------------------------ Epoch 457 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.725045
- Iter 024 / 025, Loss: 0.641151

* Train accuracy / confusion: 68.75% / [[295, 47, 11], [71, 168, 30], [50, 41, 87]], 
* Val   accuracy / confusion: 60.58% / [[33, 11, 2], [8, 22, 5], [8, 7, 8]]

------------------------------ Epoch 458 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.818635
- Iter 024 / 025, Loss: 0.555259

* Train accuracy / confusion: 69.50% / [[295, 48, 14], [70, 176, 26], [42, 44, 85]], 
* Val   accuracy / confusion: 48.08% / [[26, 18, 2], [11, 20, 4], [3, 16, 4]]

------------------------------ Epoch 459 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.660907
- Iter 024 / 025, Loss: 0.650987

* Train accuracy / confusion: 72.25% / [[306, 37, 13], [74, 178, 13], [40, 45, 94]], 
* Val   accuracy / confusion: 55.77% / [[31, 13, 2], [9, 23, 3], [7, 12, 4]]

------------------------------ Epoch 460 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.756501
- Iter 024 / 025, Loss: 0.618739

* Train accuracy / confusion: 72.38% / [[302, 42, 13], [64, 179, 23], [36, 43, 98]], 
* Val   accuracy / confusion: 60.58% / [[35, 10, 1], [9, 22, 4], [6, 11, 6]]

------------------------------ Epoch 461 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.745013
- Iter 024 / 025, Loss: 0.561073

* Train accuracy / confusion: 70.38% / [[295, 50, 14], [66, 180, 21], [48, 38, 88]], 
* Val   accuracy / confusion: 51.92% / [[31, 14, 1], [8, 19, 8], [7, 12, 4]]

------------------------------ Epoch 462 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.634545
- Iter 024 / 025, Loss: 0.632999

* Train accuracy / confusion: 71.62% / [[300, 49, 6], [77, 171, 16], [37, 42, 102]], 
* Val   accuracy / confusion: 55.77% / [[30, 12, 4], [7, 21, 7], [8, 8, 7]]

------------------------------ Epoch 463 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.804035
- Iter 024 / 025, Loss: 0.504441

* Train accuracy / confusion: 72.38% / [[301, 43, 11], [77, 169, 23], [31, 36, 109]], 
* Val   accuracy / confusion: 52.88% / [[31, 12, 3], [15, 18, 2], [10, 7, 6]]

------------------------------ Epoch 464 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.653408
- Iter 024 / 025, Loss: 0.543442

* Train accuracy / confusion: 71.88% / [[300, 47, 15], [68, 167, 28], [34, 33, 108]], 
* Val   accuracy / confusion: 49.04% / [[29, 14, 3], [9, 18, 8], [4, 15, 4]]

------------------------------ Epoch 465 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.640220
- Iter 024 / 025, Loss: 0.716979

* Train accuracy / confusion: 71.00% / [[294, 56, 9], [71, 175, 23], [36, 37, 99]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [9, 21, 5], [8, 12, 3]]

------------------------------ Epoch 466 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.668092
- Iter 024 / 025, Loss: 0.549379

* Train accuracy / confusion: 72.00% / [[309, 37, 12], [68, 174, 24], [34, 49, 93]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [9, 22, 4], [7, 13, 3]]

------------------------------ Epoch 467 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.785155
- Iter 024 / 025, Loss: 0.815072

* Train accuracy / confusion: 70.50% / [[299, 42, 15], [69, 172, 25], [39, 46, 93]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [14, 16, 5], [8, 9, 6]]

------------------------------ Epoch 468 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.707850
- Iter 024 / 025, Loss: 0.754372

* Train accuracy / confusion: 72.38% / [[308, 42, 7], [71, 175, 20], [37, 44, 96]], 
* Val   accuracy / confusion: 54.81% / [[28, 16, 2], [9, 21, 5], [8, 7, 8]]

------------------------------ Epoch 469 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.713483
- Iter 024 / 025, Loss: 0.778431

* Train accuracy / confusion: 68.25% / [[295, 51, 15], [80, 160, 26], [38, 44, 91]], 
* Val   accuracy / confusion: 57.69% / [[32, 13, 1], [10, 22, 3], [8, 9, 6]]

------------------------------ Epoch 470 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.814899
- Iter 024 / 025, Loss: 0.701662

* Train accuracy / confusion: 72.88% / [[305, 41, 10], [67, 176, 23], [30, 46, 102]], 
* Val   accuracy / confusion: 50.96% / [[29, 13, 4], [10, 18, 7], [7, 10, 6]]

------------------------------ Epoch 471 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.642511
- Iter 024 / 025, Loss: 0.661784

* Train accuracy / confusion: 71.75% / [[314, 38, 7], [69, 168, 28], [43, 41, 92]], 
* Val   accuracy / confusion: 49.04% / [[32, 12, 2], [16, 13, 6], [8, 9, 6]]

------------------------------ Epoch 472 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.729769
- Iter 024 / 025, Loss: 0.720675

* Train accuracy / confusion: 70.50% / [[296, 49, 13], [72, 170, 24], [36, 42, 98]], 
* Val   accuracy / confusion: 59.62% / [[33, 9, 4], [7, 23, 5], [9, 8, 6]]

------------------------------ Epoch 473 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.603086
- Iter 024 / 025, Loss: 0.538331

* Train accuracy / confusion: 70.38% / [[296, 48, 12], [74, 168, 26], [27, 50, 99]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [10, 17, 8], [9, 9, 5]]

------------------------------ Epoch 474 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.823748
- Iter 024 / 025, Loss: 0.682413

* Train accuracy / confusion: 72.50% / [[302, 37, 15], [70, 183, 18], [38, 42, 95]], 
* Val   accuracy / confusion: 52.88% / [[27, 13, 6], [10, 19, 6], [4, 10, 9]]

------------------------------ Epoch 475 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.632154
- Iter 024 / 025, Loss: 0.796974

* Train accuracy / confusion: 68.50% / [[301, 48, 8], [86, 156, 27], [33, 50, 91]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [8, 22, 5], [8, 10, 5]]

------------------------------ Epoch 476 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.696606
- Iter 024 / 025, Loss: 0.844378

* Train accuracy / confusion: 71.50% / [[298, 46, 12], [70, 176, 22], [34, 44, 98]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [12, 17, 6], [7, 11, 5]]

------------------------------ Epoch 477 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.532376
- Iter 024 / 025, Loss: 0.684183

* Train accuracy / confusion: 72.25% / [[307, 40, 10], [72, 178, 22], [35, 43, 93]], 
* Val   accuracy / confusion: 54.81% / [[30, 11, 5], [7, 22, 6], [5, 13, 5]]

------------------------------ Epoch 478 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.652486
- Iter 024 / 025, Loss: 0.617658

* Train accuracy / confusion: 71.00% / [[299, 41, 17], [73, 179, 18], [39, 44, 90]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [10, 16, 9], [8, 9, 6]]

------------------------------ Epoch 479 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.665762
- Iter 024 / 025, Loss: 0.821762

* Train accuracy / confusion: 69.38% / [[303, 47, 9], [79, 163, 24], [42, 44, 89]], 
* Val   accuracy / confusion: 51.92% / [[28, 14, 4], [11, 21, 3], [3, 15, 5]]

------------------------------ Epoch 480 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.753076
- Iter 024 / 025, Loss: 0.528894

* Train accuracy / confusion: 69.38% / [[290, 50, 15], [78, 165, 24], [39, 39, 100]], 
* Val   accuracy / confusion: 45.19% / [[31, 13, 2], [14, 12, 9], [8, 11, 4]]

------------------------------ Epoch 481 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.565913
- Iter 024 / 025, Loss: 0.689273

* Train accuracy / confusion: 73.00% / [[305, 39, 13], [59, 184, 25], [39, 41, 95]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [8, 23, 4], [6, 14, 3]]

------------------------------ Epoch 482 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.694061
- Iter 024 / 025, Loss: 0.593939

* Train accuracy / confusion: 70.50% / [[290, 50, 11], [71, 177, 23], [35, 46, 97]], 
* Val   accuracy / confusion: 57.69% / [[37, 7, 2], [10, 19, 6], [7, 12, 4]]

------------------------------ Epoch 483 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.787372
- Iter 024 / 025, Loss: 0.649835

* Train accuracy / confusion: 72.75% / [[304, 38, 10], [80, 175, 17], [34, 39, 103]], 
* Val   accuracy / confusion: 53.85% / [[32, 14, 0], [10, 19, 6], [6, 12, 5]]

------------------------------ Epoch 484 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.647643
- Iter 024 / 025, Loss: 0.632056

* Train accuracy / confusion: 69.75% / [[298, 49, 6], [79, 166, 28], [38, 42, 94]], 
* Val   accuracy / confusion: 55.77% / [[31, 14, 1], [12, 20, 3], [6, 10, 7]]

------------------------------ Epoch 485 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.646304
- Iter 024 / 025, Loss: 0.632314

* Train accuracy / confusion: 73.00% / [[307, 41, 13], [66, 175, 23], [35, 38, 102]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [12, 18, 5], [5, 10, 8]]

------------------------------ Epoch 486 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.743918
- Iter 024 / 025, Loss: 0.823649

* Train accuracy / confusion: 71.62% / [[307, 40, 9], [77, 172, 20], [32, 49, 94]], 
* Val   accuracy / confusion: 51.92% / [[29, 13, 4], [10, 18, 7], [4, 12, 7]]

------------------------------ Epoch 487 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.912053
- Iter 024 / 025, Loss: 0.717715

* Train accuracy / confusion: 70.62% / [[290, 50, 17], [73, 177, 19], [37, 39, 98]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [11, 21, 3], [9, 12, 2]]

------------------------------ Epoch 488 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.720448
- Iter 024 / 025, Loss: 0.639827

* Train accuracy / confusion: 69.12% / [[290, 55, 12], [73, 175, 20], [36, 51, 88]], 
* Val   accuracy / confusion: 56.73% / [[30, 13, 3], [9, 21, 5], [5, 10, 8]]

------------------------------ Epoch 489 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.648357
- Iter 024 / 025, Loss: 0.564418

* Train accuracy / confusion: 72.00% / [[296, 44, 13], [57, 187, 25], [35, 50, 93]], 
* Val   accuracy / confusion: 50.96% / [[30, 14, 2], [13, 19, 3], [5, 14, 4]]

------------------------------ Epoch 490 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.723208
- Iter 024 / 025, Loss: 0.587958

* Train accuracy / confusion: 72.50% / [[316, 36, 8], [76, 166, 24], [29, 47, 98]], 
* Val   accuracy / confusion: 57.69% / [[30, 14, 2], [10, 23, 2], [3, 13, 7]]

------------------------------ Epoch 491 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.752189
- Iter 024 / 025, Loss: 0.686854

* Train accuracy / confusion: 71.12% / [[297, 40, 19], [73, 174, 22], [33, 44, 98]], 
* Val   accuracy / confusion: 52.88% / [[29, 13, 4], [7, 20, 8], [7, 10, 6]]

------------------------------ Epoch 492 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.555636
- Iter 024 / 025, Loss: 0.556394

* Train accuracy / confusion: 72.88% / [[305, 46, 11], [57, 179, 25], [38, 40, 99]], 
* Val   accuracy / confusion: 53.85% / [[31, 13, 2], [11, 18, 6], [2, 14, 7]]

------------------------------ Epoch 493 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.590855
- Iter 024 / 025, Loss: 0.593481

* Train accuracy / confusion: 72.25% / [[303, 44, 6], [68, 185, 14], [33, 57, 90]], 
* Val   accuracy / confusion: 57.69% / [[31, 14, 1], [7, 25, 3], [5, 14, 4]]

------------------------------ Epoch 494 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.753702
- Iter 024 / 025, Loss: 0.688422

* Train accuracy / confusion: 71.12% / [[306, 41, 12], [70, 182, 19], [35, 54, 81]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [9, 19, 7], [9, 9, 5]]

------------------------------ Epoch 495 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.512935
- Iter 024 / 025, Loss: 0.567605

* Train accuracy / confusion: 72.62% / [[300, 39, 14], [66, 183, 21], [38, 41, 98]], 
* Val   accuracy / confusion: 52.88% / [[30, 13, 3], [9, 20, 6], [6, 12, 5]]

------------------------------ Epoch 496 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.795123
- Iter 024 / 025, Loss: 0.832436

* Train accuracy / confusion: 69.50% / [[295, 48, 14], [79, 168, 20], [36, 47, 93]], 
* Val   accuracy / confusion: 54.81% / [[30, 12, 4], [6, 23, 6], [6, 13, 4]]

------------------------------ Epoch 497 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.527374
- Iter 024 / 025, Loss: 0.623861

* Train accuracy / confusion: 70.62% / [[301, 43, 11], [76, 171, 25], [32, 48, 93]], 
* Val   accuracy / confusion: 52.88% / [[28, 14, 4], [11, 18, 6], [6, 8, 9]]

------------------------------ Epoch 498 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.698645
- Iter 024 / 025, Loss: 0.431435

* Train accuracy / confusion: 72.62% / [[300, 45, 14], [67, 187, 13], [36, 44, 94]], 
* Val   accuracy / confusion: 56.73% / [[29, 15, 2], [9, 23, 3], [6, 10, 7]]

------------------------------ Epoch 499 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.804095
- Iter 024 / 025, Loss: 0.524733

* Train accuracy / confusion: 71.75% / [[298, 45, 13], [57, 186, 23], [34, 54, 90]], 
* Val   accuracy / confusion: 50.00% / [[28, 16, 2], [11, 21, 3], [7, 13, 3]]

------------------------------ Epoch 500 / 500, Learning rate: 1.40e-04 ------------------------------
- Iter 012 / 025, Loss: 0.686598
- Iter 024 / 025, Loss: 0.686910

* Train accuracy / confusion: 73.12% / [[301, 44, 14], [66, 185, 18], [36, 37, 99]], 
* Val   accuracy / confusion: 58.65% / [[30, 14, 2], [3, 26, 6], [3, 15, 5]]

**************************************** Training Ends ****************************************
- Test accuracy: 51.75%

- Confusion matrix:
 [[852 452  76]
 [249 567 204]
 [112 398 180]]

- Debug table:
{   '00941': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 3, 0]},
    '00408': {'GT': 0, 'Acc': ' 13.33%', 'Pred': [4, 26, 0]},
    '00871': {'GT': 0, 'Acc': ' 96.67%', 'Pred': [29, 1, 0]},
    '00852': {'GT': 0, 'Acc': ' 83.33%', 'Pred': [25, 5, 0]},
    '00816': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00672': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00414': {'GT': 2, 'Acc': ' 50.00%', 'Pred': [0, 15, 15]},
    '00767': {'GT': 1, 'Acc': ' 83.33%', 'Pred': [4, 25, 1]},
    '01148': {'GT': 1, 'Acc': ' 26.67%', 'Pred': [19, 8, 3]},
    '00743': {'GT': 0, 'Acc': ' 26.67%', 'Pred': [8, 22, 0]},
    '00305': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '01305': {'GT': 1, 'Acc': ' 76.67%', 'Pred': [3, 23, 4]},
    '00455': {'GT': 1, 'Acc': ' 30.00%', 'Pred': [5, 9, 16]},
    '01140': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00649': {'GT': 2, 'Acc': ' 30.00%', 'Pred': [1, 20, 9]},
    '01068': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00176': {'GT': 0, 'Acc': ' 66.67%', 'Pred': [20, 10, 0]},
    '00861': {'GT': 2, 'Acc': '  3.33%', 'Pred': [1, 28, 1]},
    '00302': {'GT': 1, 'Acc': ' 50.00%', 'Pred': [3, 15, 12]},
    '00756': {'GT': 1, 'Acc': ' 56.67%', 'Pred': [13, 17, 0]},
    '00706': {'GT': 0, 'Acc': ' 76.67%', 'Pred': [23, 7, 0]},
    '01218': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '01008': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00914': {'GT': 0, 'Acc': ' 20.00%', 'Pred': [6, 24, 0]},
    '00680': {'GT': 1, 'Acc': ' 83.33%', 'Pred': [5, 25, 0]},
    '00736': {'GT': 2, 'Acc': ' 26.67%', 'Pred': [0, 22, 8]},
    '00417': {'GT': 2, 'Acc': '  0.00%', 'Pred': [3, 27, 0]},
    '01184': {'GT': 2, 'Acc': ' 16.67%', 'Pred': [0, 25, 5]},
    '00949': {'GT': 1, 'Acc': ' 70.00%', 'Pred': [8, 21, 1]},
    '01268': {'GT': 1, 'Acc': ' 90.00%', 'Pred': [3, 27, 0]},
    '00441': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00731': {'GT': 0, 'Acc': ' 96.67%', 'Pred': [29, 1, 0]},
    '00806': {'GT': 2, 'Acc': ' 16.67%', 'Pred': [8, 17, 5]},
    '00192': {'GT': 0, 'Acc': ' 36.67%', 'Pred': [11, 19, 0]},
    '01075': {'GT': 2, 'Acc': ' 10.00%', 'Pred': [15, 12, 3]},
    '00058': {'GT': 0, 'Acc': ' 70.00%', 'Pred': [21, 2, 7]},
    '01126': {'GT': 0, 'Acc': ' 73.33%', 'Pred': [22, 8, 0]},
    '00924': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '01070': {'GT': 0, 'Acc': ' 16.67%', 'Pred': [5, 25, 0]},
    '00901': {'GT': 0, 'Acc': ' 23.33%', 'Pred': [7, 23, 0]},
    '01079': {'GT': 1, 'Acc': ' 23.33%', 'Pred': [23, 7, 0]},
    '00614': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00062': {'GT': 1, 'Acc': ' 80.00%', 'Pred': [6, 24, 0]},
    '00405': {'GT': 2, 'Acc': ' 50.00%', 'Pred': [0, 15, 15]},
    '01307': {'GT': 0, 'Acc': ' 53.33%', 'Pred': [16, 11, 3]},
    '00918': {'GT': 0, 'Acc': '  3.33%', 'Pred': [1, 11, 18]},
    '00299': {'GT': 0, 'Acc': '  3.33%', 'Pred': [1, 13, 16]},
    '01046': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01106': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00822': {'GT': 0, 'Acc': ' 96.67%', 'Pred': [29, 1, 0]},
    '01037': {'GT': 1, 'Acc': ' 46.67%', 'Pred': [2, 14, 14]},
    '00599': {'GT': 1, 'Acc': ' 70.00%', 'Pred': [9, 21, 0]},
    '00517': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00832': {'GT': 2, 'Acc': '  0.00%', 'Pred': [20, 10, 0]},
    '00498': {'GT': 1, 'Acc': ' 33.33%', 'Pred': [5, 10, 15]},
    '00280': {'GT': 2, 'Acc': ' 30.00%', 'Pred': [4, 17, 9]},
    '00623': {'GT': 2, 'Acc': ' 20.00%', 'Pred': [10, 14, 6]},
    '01038': {'GT': 1, 'Acc': ' 86.67%', 'Pred': [0, 26, 4]},
    '00534': {'GT': 0, 'Acc': ' 33.33%', 'Pred': [10, 20, 0]},
    '01199': {'GT': 2, 'Acc': ' 43.33%', 'Pred': [0, 17, 13]},
    '00886': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 29, 1]},
    '00698': {'GT': 1, 'Acc': ' 73.33%', 'Pred': [2, 22, 6]},
    '00509': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00389': {'GT': 1, 'Acc': ' 40.00%', 'Pred': [18, 12, 0]},
    '01208': {'GT': 2, 'Acc': '  6.67%', 'Pred': [5, 23, 2]},
    '00004': {'GT': 2, 'Acc': ' 30.00%', 'Pred': [1, 20, 9]},
    '01039': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '01086': {'GT': 2, 'Acc': ' 20.00%', 'Pred': [16, 8, 6]},
    '00174': {'GT': 2, 'Acc': '  0.00%', 'Pred': [19, 11, 0]},
    '00226': {'GT': 2, 'Acc': ' 53.33%', 'Pred': [0, 14, 16]},
    '00794': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00289': {'GT': 2, 'Acc': ' 93.33%', 'Pred': [0, 2, 28]},
    '01027': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00855': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00385': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 13, 17]},
    '00022': {'GT': 1, 'Acc': ' 36.67%', 'Pred': [3, 11, 16]},
    '00220': {'GT': 1, 'Acc': ' 26.67%', 'Pred': [18, 8, 4]},
    '00635': {'GT': 2, 'Acc': ' 33.33%', 'Pred': [6, 14, 10]},
    '00122': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00505': {'GT': 0, 'Acc': ' 66.67%', 'Pred': [20, 10, 0]},
    '01014': {'GT': 1, 'Acc': '  0.00%', 'Pred': [29, 0, 1]},
    '00407': {'GT': 1, 'Acc': ' 10.00%', 'Pred': [26, 3, 1]},
    '01300': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00301': {'GT': 1, 'Acc': ' 76.67%', 'Pred': [0, 23, 7]},
    '01349': {'GT': 1, 'Acc': ' 80.00%', 'Pred': [3, 24, 3]},
    '00005': {'GT': 2, 'Acc': ' 23.33%', 'Pred': [3, 20, 7]},
    '01169': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 27, 3]},
    '00124': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00243': {'GT': 1, 'Acc': '  6.67%', 'Pred': [24, 2, 4]},
    '01240': {'GT': 1, 'Acc': ' 13.33%', 'Pred': [0, 4, 26]},
    '00589': {'GT': 0, 'Acc': ' 66.67%', 'Pred': [20, 2, 8]},
    '01172': {'GT': 1, 'Acc': ' 46.67%', 'Pred': [0, 14, 16]},
    '01156': {'GT': 1, 'Acc': ' 80.00%', 'Pred': [6, 24, 0]},
    '00329': {'GT': 0, 'Acc': ' 56.67%', 'Pred': [17, 13, 0]},
    '01333': {'GT': 0, 'Acc': ' 23.33%', 'Pred': [7, 23, 0]},
    '01364': {'GT': 1, 'Acc': ' 33.33%', 'Pred': [0, 10, 20]},
    '00415': {'GT': 1, 'Acc': ' 60.00%', 'Pred': [12, 18, 0]},
    '00603': {'GT': 2, 'Acc': ' 43.33%', 'Pred': [0, 17, 13]},
    '00096': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01250': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00125': {'GT': 0, 'Acc': ' 46.67%', 'Pred': [14, 13, 3]},
    '00447': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00820': {'GT': 1, 'Acc': '  0.00%', 'Pred': [0, 0, 30]}}

1D CNN: Siamese model2

class SiameseNet2(nn.Module):
    def __init__(self, n_input=20, n_output=3, n_channel=64,
                 use_age=True, final_pool='average'):
        super().__init__()
        
        if final_pool not in {'average', 'max'}:
            raise ValueError("final_pool must be set to one of ['average', 'max']")
        
        self.use_age = use_age
        
        self.conv1 = nn.Conv1d(n_input, n_channel, kernel_size=25, dilation=2, stride=3)
        self.bn1 = nn.BatchNorm1d(n_channel)
        self.pool1 = nn.MaxPool1d(3)
        
        self.conv2 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn2 = nn.BatchNorm1d(n_channel)
        self.pool2 = nn.MaxPool1d(2)
        
        self.conv3 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn3 = nn.BatchNorm1d(n_channel)
        self.pool3 = nn.MaxPool1d(2)
        
        self.conv4 = nn.Conv1d(n_channel, n_channel*2, kernel_size=7)
        self.bn4 = nn.BatchNorm1d(n_channel*2)
        
        if final_pool == 'average':
            self.final_pool = nn.AdaptiveAvgPool1d(1)
        elif final_pool == 'max':
            self.final_pool = nn.AdaptiveMaxPool1d(1)

        if self.use_age:        
            self.fc1_1 = nn.Linear(n_channel*2 + 1, n_channel)
            self.fc1_2 = nn.Linear(n_channel*2 + 1, n_channel)
        else:
            self.fc1_1 = nn.Linear(n_channel*2, n_channel)
            self.fc1_2 = nn.Linear(n_channel*2, n_channel)
        self.bnfc1_1 = nn.BatchNorm1d(n_channel)
        self.bnfc1_2 = nn.BatchNorm1d(n_channel)
                        
        self.fc2 = nn.Linear(n_channel*2, n_channel)
        self.bnfc2 = nn.BatchNorm1d(n_channel)

        self.fc3 = nn.Linear(n_channel, n_channel//2)
        self.bnfc3 = nn.BatchNorm1d(n_channel//2)
        
        self.fc4 = nn.Linear(n_channel//2, n_output)
        
    def reset_weights(self):
        for m in self.modules():
            if hasattr(m, 'reset_parameters'):
                m.reset_parameters()

    def forward(self, x, age):
        # conv-bn-relu-pool 
        x = self.conv1(x)
        x = F.relu(self.bn1(x))
        x = self.pool1(x)

        x = self.conv2(x)
        x = F.relu(self.bn2(x))
        x = self.pool2(x)

        x = self.conv3(x)
        x = F.relu(self.bn3(x))
        x = self.pool3(x)

        x = self.conv4(x)
        x = F.relu(self.bn4(x))
        
        L = x.shape[2]
        x1 = x[:, :, :L//2]
        x2 = x[:, :, L//2:]
        
        x1 = self.final_pool(x1).squeeze()
        x2 = self.final_pool(x2).squeeze()
        
        if self.use_age:
            x1 = torch.cat((x1, age.reshape(-1, 1)), dim=1)
            x2 = torch.cat((x2, age.reshape(-1, 1)), dim=1)

        # fc-bn-relu-fc
        x1 = self.fc1_1(x1)
        x1 = self.bnfc1_1(x1)
        x1 = F.relu(x1)
        
        x2 = self.fc1_2(x2)
        x2 = self.bnfc1_2(x2)
        x2 = F.relu(x2)
        
        x = torch.cat((x1, x2), dim=1)

        x = self.fc2(x)
        x = self.bnfc2(x)
        x = F.relu(x)

        x = self.fc3(x)
        x = self.bnfc3(x)
        x = F.relu(x)
        
        x = self.fc4(x)
        
        return x
        # return F.log_softmax(x, dim=1)

model = SiameseNet2(n_input=train_dataset[0]['signal'].shape[0], 
                    n_output=3, 
                    use_age=True, 
                    final_pool='average')

model = model.to(device, dtype=torch.float32)
print(model)
print()

# tensorboard visualization
visualize_network_tensorboard(model, 'SiameseNet_2')

# number of parameters
n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

SiameseNet2(
  (conv1): Conv1d(20, 64, kernel_size=(25,), stride=(3,), dilation=(2,))
  (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool1): MaxPool1d(kernel_size=3, stride=3, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv3): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn3): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool3): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv4): Conv1d(64, 128, kernel_size=(7,), stride=(1,))
  (bn4): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (final_pool): AdaptiveAvgPool1d(output_size=1)
  (fc1_1): Linear(in_features=129, out_features=64, bias=True)
  (fc1_2): Linear(in_features=129, out_features=64, bias=True)
  (bnfc1_1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (bnfc1_2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc2): Linear(in_features=128, out_features=64, bias=True)
  (bnfc2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc3): Linear(in_features=64, out_features=32, bias=True)
  (bnfc3): BatchNorm1d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc4): Linear(in_features=32, out_features=3, bias=True)
)

The Number of parameters of the model: 175,171

record = learning_rate_search(model, 
                              min_log_lr=-5.0, 
                              max_log_lr=-1.0, 
                              trials=500, 
                              epochs=1)
draw_learning_rate_record(record)
best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

# best_log_lr = -1.8528783104568398
print('best_log_lr:', best_log_lr)

best_log_lr: -2.4187821476492295

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 200
log_interval = len(train_loader) // 2

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d}, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e} {"-"*30}')
    
    # train
    loss, train_accuracy, train_confusion = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy, val_confusion = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train accuracy / confusion: {train_accuracy:.2f}% / {train_confusion.tolist()}, ')
    print(f'* Val   accuracy / confusion: {val_accuracy:.2f}% / {val_confusion.tolist()}')
    print()
    
# test
test_accuracy, test_confusion, test_debug = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print()
print('- Confusion matrix:\n', test_confusion)
print()
print('- Debug table:')
pprint.pp(test_debug, indent=4)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(test_confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.082891
- Iter 024 / 025, Loss: 0.893879

* Train accuracy / confusion: 46.25% / [[241, 65, 48], [129, 71, 69], [64, 55, 58]], 
* Val   accuracy / confusion: 53.85% / [[33, 13, 0], [13, 22, 0], [10, 12, 1]]

------------------------------ Epoch 002 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.989449
- Iter 024 / 025, Loss: 0.917536

* Train accuracy / confusion: 51.62% / [[265, 87, 6], [122, 128, 17], [55, 100, 20]], 
* Val   accuracy / confusion: 44.23% / [[33, 8, 5], [13, 9, 13], [8, 11, 4]]

------------------------------ Epoch 003 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.058867
- Iter 024 / 025, Loss: 1.101479

* Train accuracy / confusion: 53.25% / [[265, 75, 19], [106, 119, 38], [64, 72, 42]], 
* Val   accuracy / confusion: 54.81% / [[31, 14, 1], [5, 26, 4], [7, 16, 0]]

------------------------------ Epoch 004 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.062993
- Iter 024 / 025, Loss: 0.919948

* Train accuracy / confusion: 51.00% / [[250, 77, 26], [112, 110, 48], [54, 75, 48]], 
* Val   accuracy / confusion: 54.81% / [[37, 7, 2], [12, 18, 5], [9, 12, 2]]

------------------------------ Epoch 005 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.034611
- Iter 024 / 025, Loss: 0.920608

* Train accuracy / confusion: 51.00% / [[250, 82, 24], [116, 112, 40], [57, 73, 46]], 
* Val   accuracy / confusion: 50.96% / [[28, 13, 5], [3, 21, 11], [4, 15, 4]]

------------------------------ Epoch 006 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.816769
- Iter 024 / 025, Loss: 0.954646

* Train accuracy / confusion: 52.38% / [[248, 98, 7], [97, 155, 21], [47, 111, 16]], 
* Val   accuracy / confusion: 53.85% / [[34, 12, 0], [13, 22, 0], [6, 17, 0]]

------------------------------ Epoch 007 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.968863
- Iter 024 / 025, Loss: 0.979768

* Train accuracy / confusion: 53.88% / [[258, 91, 10], [110, 119, 39], [59, 60, 54]], 
* Val   accuracy / confusion: 54.81% / [[34, 9, 3], [8, 20, 7], [5, 15, 3]]

------------------------------ Epoch 008 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.939378
- Iter 024 / 025, Loss: 0.953570

* Train accuracy / confusion: 53.38% / [[245, 98, 14], [98, 133, 42], [47, 74, 49]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [8, 22, 5], [6, 16, 1]]

------------------------------ Epoch 009 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.968747
- Iter 024 / 025, Loss: 0.819025

* Train accuracy / confusion: 52.88% / [[255, 90, 10], [104, 139, 26], [48, 99, 29]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 010 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.009406
- Iter 024 / 025, Loss: 1.116933

* Train accuracy / confusion: 53.75% / [[247, 86, 20], [97, 131, 42], [54, 71, 52]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [7, 18, 10], [6, 12, 5]]

------------------------------ Epoch 011 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.899658
- Iter 024 / 025, Loss: 0.871321

* Train accuracy / confusion: 54.00% / [[253, 78, 26], [104, 123, 40], [51, 69, 56]], 
* Val   accuracy / confusion: 53.85% / [[32, 13, 1], [8, 22, 5], [6, 15, 2]]

------------------------------ Epoch 012 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.950913
- Iter 024 / 025, Loss: 1.062857

* Train accuracy / confusion: 54.38% / [[246, 83, 24], [104, 124, 41], [55, 58, 65]], 
* Val   accuracy / confusion: 52.88% / [[31, 10, 5], [6, 20, 9], [6, 13, 4]]

------------------------------ Epoch 013 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.023990
- Iter 024 / 025, Loss: 1.012303

* Train accuracy / confusion: 53.62% / [[245, 92, 17], [98, 149, 25], [49, 90, 35]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [8, 23, 4], [6, 15, 2]]

------------------------------ Epoch 014 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.866856
- Iter 024 / 025, Loss: 1.022563

* Train accuracy / confusion: 54.12% / [[240, 94, 24], [86, 138, 43], [49, 71, 55]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 015 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.800208
- Iter 024 / 025, Loss: 0.833786

* Train accuracy / confusion: 53.38% / [[240, 108, 7], [95, 155, 20], [50, 93, 32]], 
* Val   accuracy / confusion: 57.69% / [[32, 12, 2], [5, 27, 3], [5, 17, 1]]

------------------------------ Epoch 016 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.932383
- Iter 024 / 025, Loss: 0.803389

* Train accuracy / confusion: 54.62% / [[253, 83, 25], [101, 131, 32], [45, 77, 53]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [8, 23, 4], [4, 18, 1]]

------------------------------ Epoch 017 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.904520
- Iter 024 / 025, Loss: 0.960148

* Train accuracy / confusion: 51.88% / [[245, 84, 23], [108, 116, 47], [54, 69, 54]], 
* Val   accuracy / confusion: 56.73% / [[34, 12, 0], [7, 24, 4], [5, 17, 1]]

------------------------------ Epoch 018 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.851942
- Iter 024 / 025, Loss: 1.184780

* Train accuracy / confusion: 53.00% / [[224, 120, 14], [79, 150, 36], [48, 79, 50]], 
* Val   accuracy / confusion: 52.88% / [[27, 15, 4], [4, 25, 6], [4, 16, 3]]

------------------------------ Epoch 019 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.990958
- Iter 024 / 025, Loss: 1.088246

* Train accuracy / confusion: 52.50% / [[252, 88, 19], [111, 109, 45], [58, 59, 59]], 
* Val   accuracy / confusion: 50.96% / [[32, 10, 4], [7, 18, 10], [6, 14, 3]]

------------------------------ Epoch 020 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.991693
- Iter 024 / 025, Loss: 0.872186

* Train accuracy / confusion: 52.75% / [[261, 85, 9], [107, 130, 33], [58, 86, 31]], 
* Val   accuracy / confusion: 54.81% / [[31, 15, 0], [7, 25, 3], [6, 16, 1]]

------------------------------ Epoch 021 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.044638
- Iter 024 / 025, Loss: 0.772852

* Train accuracy / confusion: 55.62% / [[260, 76, 14], [116, 126, 32], [60, 57, 59]], 
* Val   accuracy / confusion: 50.96% / [[33, 9, 4], [7, 18, 10], [6, 15, 2]]

------------------------------ Epoch 022 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.230562
- Iter 024 / 025, Loss: 0.877943

* Train accuracy / confusion: 54.00% / [[261, 75, 18], [115, 113, 40], [60, 60, 58]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [8, 21, 6], [6, 16, 1]]

------------------------------ Epoch 023 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.081461
- Iter 024 / 025, Loss: 0.929778

* Train accuracy / confusion: 54.88% / [[243, 98, 14], [95, 139, 34], [52, 68, 57]], 
* Val   accuracy / confusion: 54.81% / [[29, 13, 4], [4, 26, 5], [4, 17, 2]]

------------------------------ Epoch 024 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.008793
- Iter 024 / 025, Loss: 0.878009

* Train accuracy / confusion: 53.75% / [[242, 99, 17], [98, 134, 36], [49, 71, 54]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [8, 21, 6], [5, 14, 4]]

------------------------------ Epoch 025 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.875834
- Iter 024 / 025, Loss: 0.871902

* Train accuracy / confusion: 54.00% / [[247, 90, 19], [105, 128, 35], [50, 69, 57]], 
* Val   accuracy / confusion: 57.69% / [[33, 10, 3], [6, 24, 5], [7, 13, 3]]

------------------------------ Epoch 026 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.927357
- Iter 024 / 025, Loss: 0.996658

* Train accuracy / confusion: 55.12% / [[227, 104, 22], [73, 160, 35], [42, 83, 54]], 
* Val   accuracy / confusion: 53.85% / [[29, 17, 0], [5, 25, 5], [5, 16, 2]]

------------------------------ Epoch 027 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.951165
- Iter 024 / 025, Loss: 0.843568

* Train accuracy / confusion: 54.00% / [[248, 94, 14], [95, 140, 33], [56, 76, 44]], 
* Val   accuracy / confusion: 54.81% / [[30, 15, 1], [5, 25, 5], [5, 16, 2]]

------------------------------ Epoch 028 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.952282
- Iter 024 / 025, Loss: 0.825571

* Train accuracy / confusion: 54.38% / [[255, 83, 20], [107, 119, 41], [51, 63, 61]], 
* Val   accuracy / confusion: 52.88% / [[29, 15, 2], [6, 24, 5], [5, 16, 2]]

------------------------------ Epoch 029 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.892155
- Iter 024 / 025, Loss: 0.952548

* Train accuracy / confusion: 57.38% / [[235, 99, 22], [84, 157, 29], [45, 62, 67]], 
* Val   accuracy / confusion: 50.96% / [[24, 19, 3], [3, 26, 6], [4, 16, 3]]

------------------------------ Epoch 030 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.898986
- Iter 024 / 025, Loss: 0.977184

* Train accuracy / confusion: 53.12% / [[237, 98, 21], [97, 123, 45], [48, 66, 65]], 
* Val   accuracy / confusion: 50.96% / [[27, 17, 2], [5, 25, 5], [4, 18, 1]]

------------------------------ Epoch 031 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.981139
- Iter 024 / 025, Loss: 0.996640

* Train accuracy / confusion: 55.50% / [[234, 102, 19], [78, 156, 37], [41, 79, 54]], 
* Val   accuracy / confusion: 51.92% / [[29, 14, 3], [4, 24, 7], [4, 18, 1]]

------------------------------ Epoch 032 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.889103
- Iter 024 / 025, Loss: 0.965924

* Train accuracy / confusion: 56.38% / [[257, 83, 15], [101, 134, 33], [62, 55, 60]], 
* Val   accuracy / confusion: 54.81% / [[32, 9, 5], [6, 21, 8], [6, 13, 4]]

------------------------------ Epoch 033 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.001593
- Iter 024 / 025, Loss: 0.918627

* Train accuracy / confusion: 55.12% / [[250, 84, 25], [97, 118, 52], [46, 55, 73]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [5, 23, 7], [5, 16, 2]]

------------------------------ Epoch 034 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.096561
- Iter 024 / 025, Loss: 0.947948

* Train accuracy / confusion: 53.38% / [[234, 87, 30], [100, 127, 43], [46, 67, 66]], 
* Val   accuracy / confusion: 50.00% / [[30, 11, 5], [5, 21, 9], [4, 18, 1]]

------------------------------ Epoch 035 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.014882
- Iter 024 / 025, Loss: 0.972882

* Train accuracy / confusion: 54.88% / [[237, 103, 18], [91, 144, 37], [48, 64, 58]], 
* Val   accuracy / confusion: 55.77% / [[30, 14, 2], [4, 25, 6], [6, 14, 3]]

------------------------------ Epoch 036 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.924145
- Iter 024 / 025, Loss: 1.058652

* Train accuracy / confusion: 56.75% / [[253, 94, 13], [88, 143, 36], [51, 64, 58]], 
* Val   accuracy / confusion: 51.92% / [[28, 16, 2], [3, 24, 8], [3, 18, 2]]

------------------------------ Epoch 037 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.006181
- Iter 024 / 025, Loss: 1.010145

* Train accuracy / confusion: 53.62% / [[250, 84, 18], [106, 118, 45], [50, 68, 61]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [9, 19, 7], [4, 15, 4]]

------------------------------ Epoch 038 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.889554
- Iter 024 / 025, Loss: 0.927113

* Train accuracy / confusion: 54.38% / [[253, 85, 18], [107, 125, 37], [59, 59, 57]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [6, 24, 5], [4, 17, 2]]

------------------------------ Epoch 039 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.999139
- Iter 024 / 025, Loss: 1.002435

* Train accuracy / confusion: 53.38% / [[224, 117, 11], [87, 165, 20], [45, 93, 38]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [5, 22, 8], [4, 16, 3]]

------------------------------ Epoch 040 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.010933
- Iter 024 / 025, Loss: 0.890130

* Train accuracy / confusion: 53.88% / [[240, 91, 24], [93, 123, 54], [60, 47, 68]], 
* Val   accuracy / confusion: 58.65% / [[31, 14, 1], [2, 28, 5], [5, 16, 2]]

------------------------------ Epoch 041 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.858243
- Iter 024 / 025, Loss: 0.917950

* Train accuracy / confusion: 53.12% / [[230, 105, 18], [90, 156, 26], [44, 92, 39]], 
* Val   accuracy / confusion: 52.88% / [[28, 14, 4], [4, 24, 7], [5, 15, 3]]

------------------------------ Epoch 042 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.886928
- Iter 024 / 025, Loss: 0.849516

* Train accuracy / confusion: 55.25% / [[249, 92, 14], [94, 130, 47], [51, 60, 63]], 
* Val   accuracy / confusion: 57.69% / [[34, 7, 5], [6, 22, 7], [8, 11, 4]]

------------------------------ Epoch 043 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.828186
- Iter 024 / 025, Loss: 0.852088

* Train accuracy / confusion: 55.88% / [[261, 77, 21], [95, 126, 46], [53, 61, 60]], 
* Val   accuracy / confusion: 58.65% / [[34, 7, 5], [4, 25, 6], [5, 16, 2]]

------------------------------ Epoch 044 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.849835
- Iter 024 / 025, Loss: 0.938411

* Train accuracy / confusion: 54.88% / [[247, 91, 16], [99, 130, 42], [57, 56, 62]], 
* Val   accuracy / confusion: 56.73% / [[34, 10, 2], [5, 23, 7], [6, 15, 2]]

------------------------------ Epoch 045 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.994493
- Iter 024 / 025, Loss: 0.928033

* Train accuracy / confusion: 54.00% / [[260, 70, 25], [116, 105, 49], [52, 56, 67]], 
* Val   accuracy / confusion: 48.08% / [[30, 10, 6], [7, 14, 14], [5, 12, 6]]

------------------------------ Epoch 046 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.913559
- Iter 024 / 025, Loss: 0.911697

* Train accuracy / confusion: 53.50% / [[232, 96, 28], [83, 129, 54], [38, 73, 67]], 
* Val   accuracy / confusion: 54.81% / [[29, 12, 5], [5, 23, 7], [4, 14, 5]]

------------------------------ Epoch 047 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.894316
- Iter 024 / 025, Loss: 0.887538

* Train accuracy / confusion: 57.50% / [[243, 82, 30], [79, 137, 54], [45, 50, 80]], 
* Val   accuracy / confusion: 52.88% / [[30, 10, 6], [2, 19, 14], [5, 12, 6]]

------------------------------ Epoch 048 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.797370
- Iter 024 / 025, Loss: 0.884948

* Train accuracy / confusion: 55.62% / [[245, 81, 27], [91, 124, 55], [46, 55, 76]], 
* Val   accuracy / confusion: 50.00% / [[25, 15, 6], [3, 21, 11], [2, 15, 6]]

------------------------------ Epoch 049 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.860876
- Iter 024 / 025, Loss: 1.012764

* Train accuracy / confusion: 54.62% / [[236, 101, 20], [87, 129, 54], [43, 58, 72]], 
* Val   accuracy / confusion: 57.69% / [[31, 13, 2], [2, 24, 9], [3, 15, 5]]

------------------------------ Epoch 050 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.880876
- Iter 024 / 025, Loss: 1.011590

* Train accuracy / confusion: 58.88% / [[245, 82, 27], [76, 148, 44], [46, 54, 78]], 
* Val   accuracy / confusion: 55.77% / [[30, 13, 3], [5, 23, 7], [5, 13, 5]]

------------------------------ Epoch 051 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.859172
- Iter 024 / 025, Loss: 0.986074

* Train accuracy / confusion: 55.12% / [[241, 102, 14], [87, 139, 46], [47, 63, 61]], 
* Val   accuracy / confusion: 48.08% / [[27, 15, 4], [5, 21, 9], [6, 15, 2]]

------------------------------ Epoch 052 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.117880
- Iter 024 / 025, Loss: 1.159732

* Train accuracy / confusion: 57.75% / [[250, 76, 30], [87, 136, 42], [48, 55, 76]], 
* Val   accuracy / confusion: 51.92% / [[30, 12, 4], [5, 18, 12], [5, 12, 6]]

------------------------------ Epoch 053 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.874022
- Iter 024 / 025, Loss: 0.981475

* Train accuracy / confusion: 55.75% / [[256, 83, 20], [91, 132, 41], [54, 65, 58]], 
* Val   accuracy / confusion: 60.58% / [[34, 6, 6], [4, 25, 6], [5, 14, 4]]

------------------------------ Epoch 054 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.886868
- Iter 024 / 025, Loss: 0.835706

* Train accuracy / confusion: 55.88% / [[257, 78, 18], [102, 125, 44], [54, 57, 65]], 
* Val   accuracy / confusion: 52.88% / [[32, 8, 6], [6, 18, 11], [4, 14, 5]]

------------------------------ Epoch 055 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.152983
- Iter 024 / 025, Loss: 0.836151

* Train accuracy / confusion: 59.50% / [[251, 87, 20], [79, 151, 38], [36, 64, 74]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [9, 18, 8], [6, 13, 4]]

------------------------------ Epoch 056 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.867853
- Iter 024 / 025, Loss: 1.094792

* Train accuracy / confusion: 54.50% / [[247, 85, 23], [98, 122, 50], [44, 64, 67]], 
* Val   accuracy / confusion: 57.69% / [[32, 12, 2], [5, 26, 4], [6, 15, 2]]

------------------------------ Epoch 057 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.921465
- Iter 024 / 025, Loss: 1.100702

* Train accuracy / confusion: 55.62% / [[259, 82, 17], [93, 133, 42], [45, 76, 53]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [5, 21, 9], [8, 11, 4]]

------------------------------ Epoch 058 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.892515
- Iter 024 / 025, Loss: 0.913486

* Train accuracy / confusion: 55.50% / [[248, 86, 20], [92, 128, 50], [50, 58, 68]], 
* Val   accuracy / confusion: 52.88% / [[29, 12, 5], [3, 21, 11], [4, 14, 5]]

------------------------------ Epoch 059 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.031762
- Iter 024 / 025, Loss: 0.847558

* Train accuracy / confusion: 56.12% / [[248, 78, 25], [92, 127, 53], [47, 56, 74]], 
* Val   accuracy / confusion: 54.81% / [[36, 6, 4], [9, 15, 11], [8, 9, 6]]

------------------------------ Epoch 060 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.985832
- Iter 024 / 025, Loss: 0.890094

* Train accuracy / confusion: 57.38% / [[257, 75, 24], [94, 121, 56], [50, 42, 81]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [6, 21, 8], [6, 13, 4]]

------------------------------ Epoch 061 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.931296
- Iter 024 / 025, Loss: 0.984770

* Train accuracy / confusion: 57.62% / [[249, 94, 13], [80, 148, 43], [42, 67, 64]], 
* Val   accuracy / confusion: 54.81% / [[31, 11, 4], [6, 21, 8], [7, 11, 5]]

------------------------------ Epoch 062 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.919109
- Iter 024 / 025, Loss: 0.975481

* Train accuracy / confusion: 55.62% / [[262, 70, 26], [95, 120, 52], [55, 57, 63]], 
* Val   accuracy / confusion: 52.88% / [[33, 8, 5], [7, 18, 10], [6, 13, 4]]

------------------------------ Epoch 063 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.869461
- Iter 024 / 025, Loss: 0.938032

* Train accuracy / confusion: 57.25% / [[248, 84, 24], [94, 128, 43], [45, 52, 82]], 
* Val   accuracy / confusion: 54.81% / [[34, 9, 3], [8, 18, 9], [2, 16, 5]]

------------------------------ Epoch 064 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.766731
- Iter 024 / 025, Loss: 0.736555

* Train accuracy / confusion: 56.88% / [[257, 77, 23], [88, 127, 52], [49, 56, 71]], 
* Val   accuracy / confusion: 49.04% / [[28, 11, 7], [4, 17, 14], [6, 11, 6]]

------------------------------ Epoch 065 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.876165
- Iter 024 / 025, Loss: 0.799880

* Train accuracy / confusion: 58.88% / [[277, 65, 16], [105, 118, 45], [60, 38, 76]], 
* Val   accuracy / confusion: 51.92% / [[32, 8, 6], [7, 17, 11], [10, 8, 5]]

------------------------------ Epoch 066 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.932295
- Iter 024 / 025, Loss: 0.741623

* Train accuracy / confusion: 58.88% / [[259, 70, 27], [94, 131, 44], [41, 53, 81]], 
* Val   accuracy / confusion: 53.85% / [[30, 13, 3], [4, 22, 9], [4, 15, 4]]

------------------------------ Epoch 067 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.670973
- Iter 024 / 025, Loss: 0.891851

* Train accuracy / confusion: 56.50% / [[257, 83, 17], [94, 130, 42], [52, 60, 65]], 
* Val   accuracy / confusion: 51.92% / [[34, 9, 3], [7, 17, 11], [9, 11, 3]]

------------------------------ Epoch 068 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.878840
- Iter 024 / 025, Loss: 0.855464

* Train accuracy / confusion: 57.38% / [[276, 69, 17], [98, 120, 46], [56, 55, 63]], 
* Val   accuracy / confusion: 51.92% / [[31, 10, 5], [10, 20, 5], [8, 12, 3]]

------------------------------ Epoch 069 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.865961
- Iter 024 / 025, Loss: 0.897243

* Train accuracy / confusion: 58.38% / [[263, 74, 17], [94, 123, 50], [48, 50, 81]], 
* Val   accuracy / confusion: 58.65% / [[32, 11, 3], [3, 24, 8], [7, 11, 5]]

------------------------------ Epoch 070 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.842991
- Iter 024 / 025, Loss: 0.874062

* Train accuracy / confusion: 56.75% / [[260, 77, 21], [104, 120, 44], [54, 46, 74]], 
* Val   accuracy / confusion: 52.88% / [[32, 8, 6], [7, 16, 12], [6, 10, 7]]

------------------------------ Epoch 071 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.865064
- Iter 024 / 025, Loss: 0.728601

* Train accuracy / confusion: 57.25% / [[255, 75, 25], [93, 115, 58], [46, 45, 88]], 
* Val   accuracy / confusion: 52.88% / [[27, 11, 8], [6, 19, 10], [2, 12, 9]]

------------------------------ Epoch 072 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.874938
- Iter 024 / 025, Loss: 0.758550

* Train accuracy / confusion: 58.62% / [[270, 74, 15], [100, 128, 40], [44, 58, 71]], 
* Val   accuracy / confusion: 53.85% / [[32, 8, 6], [4, 21, 10], [4, 16, 3]]

------------------------------ Epoch 073 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.122630
- Iter 024 / 025, Loss: 0.860219

* Train accuracy / confusion: 57.50% / [[254, 79, 23], [90, 134, 47], [44, 57, 72]], 
* Val   accuracy / confusion: 56.73% / [[34, 9, 3], [4, 23, 8], [6, 15, 2]]

------------------------------ Epoch 074 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.913522
- Iter 024 / 025, Loss: 0.765375

* Train accuracy / confusion: 57.50% / [[273, 70, 17], [106, 125, 33], [47, 67, 62]], 
* Val   accuracy / confusion: 55.77% / [[34, 7, 5], [6, 21, 8], [7, 13, 3]]

------------------------------ Epoch 075 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.813841
- Iter 024 / 025, Loss: 0.972833

* Train accuracy / confusion: 57.88% / [[255, 79, 21], [94, 127, 50], [48, 45, 81]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [9, 15, 11], [7, 10, 6]]

------------------------------ Epoch 076 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.876691
- Iter 024 / 025, Loss: 0.941060

* Train accuracy / confusion: 58.62% / [[272, 62, 24], [98, 124, 48], [47, 52, 73]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [12, 13, 10], [6, 10, 7]]

------------------------------ Epoch 077 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.863029
- Iter 024 / 025, Loss: 0.890486

* Train accuracy / confusion: 57.75% / [[252, 80, 23], [88, 134, 47], [49, 51, 76]], 
* Val   accuracy / confusion: 55.77% / [[33, 7, 6], [4, 23, 8], [7, 14, 2]]

------------------------------ Epoch 078 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.744232
- Iter 024 / 025, Loss: 0.734167

* Train accuracy / confusion: 58.88% / [[262, 73, 23], [87, 134, 46], [50, 50, 75]], 
* Val   accuracy / confusion: 54.81% / [[35, 8, 3], [5, 20, 10], [4, 17, 2]]

------------------------------ Epoch 079 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.918067
- Iter 024 / 025, Loss: 0.983902

* Train accuracy / confusion: 57.12% / [[254, 81, 18], [88, 130, 50], [45, 61, 73]], 
* Val   accuracy / confusion: 55.77% / [[31, 11, 4], [4, 22, 9], [5, 13, 5]]

------------------------------ Epoch 080 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.838129
- Iter 024 / 025, Loss: 0.780942

* Train accuracy / confusion: 60.25% / [[273, 61, 23], [90, 123, 54], [48, 42, 86]], 
* Val   accuracy / confusion: 54.81% / [[37, 5, 4], [10, 15, 10], [6, 12, 5]]

------------------------------ Epoch 081 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.933615
- Iter 024 / 025, Loss: 0.886101

* Train accuracy / confusion: 57.88% / [[258, 77, 19], [93, 121, 58], [44, 46, 84]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [6, 19, 10], [3, 15, 5]]

------------------------------ Epoch 082 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.834508
- Iter 024 / 025, Loss: 0.948723

* Train accuracy / confusion: 59.12% / [[260, 76, 17], [90, 135, 45], [53, 46, 78]], 
* Val   accuracy / confusion: 47.12% / [[31, 11, 4], [8, 16, 11], [6, 15, 2]]

------------------------------ Epoch 083 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.027786
- Iter 024 / 025, Loss: 0.869611

* Train accuracy / confusion: 60.38% / [[280, 52, 23], [98, 129, 41], [55, 48, 74]], 
* Val   accuracy / confusion: 54.81% / [[37, 5, 4], [13, 16, 6], [7, 12, 4]]

------------------------------ Epoch 084 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.973286
- Iter 024 / 025, Loss: 0.843717

* Train accuracy / confusion: 58.88% / [[266, 65, 24], [93, 133, 45], [56, 46, 72]], 
* Val   accuracy / confusion: 52.88% / [[35, 8, 3], [13, 16, 6], [9, 10, 4]]

------------------------------ Epoch 085 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.006134
- Iter 024 / 025, Loss: 0.970160

* Train accuracy / confusion: 60.75% / [[288, 58, 15], [97, 124, 44], [55, 45, 74]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [6, 18, 11], [5, 12, 6]]

------------------------------ Epoch 086 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.933229
- Iter 024 / 025, Loss: 0.875389

* Train accuracy / confusion: 61.12% / [[283, 62, 11], [89, 140, 42], [59, 48, 66]], 
* Val   accuracy / confusion: 54.81% / [[37, 7, 2], [10, 18, 7], [6, 15, 2]]

------------------------------ Epoch 087 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.941317
- Iter 024 / 025, Loss: 0.764614

* Train accuracy / confusion: 59.25% / [[276, 64, 18], [99, 123, 43], [51, 51, 75]], 
* Val   accuracy / confusion: 49.04% / [[36, 6, 4], [17, 8, 10], [8, 8, 7]]

------------------------------ Epoch 088 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.002972
- Iter 024 / 025, Loss: 1.056571

* Train accuracy / confusion: 59.75% / [[271, 66, 20], [96, 127, 46], [50, 44, 80]], 
* Val   accuracy / confusion: 49.04% / [[34, 6, 6], [16, 9, 10], [7, 8, 8]]

------------------------------ Epoch 089 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.815944
- Iter 024 / 025, Loss: 0.844690

* Train accuracy / confusion: 59.00% / [[272, 64, 18], [106, 128, 37], [54, 49, 72]], 
* Val   accuracy / confusion: 56.73% / [[37, 7, 2], [13, 13, 9], [5, 9, 9]]

------------------------------ Epoch 090 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.669833
- Iter 024 / 025, Loss: 0.746952

* Train accuracy / confusion: 58.88% / [[273, 64, 16], [101, 117, 53], [53, 42, 81]], 
* Val   accuracy / confusion: 50.96% / [[37, 5, 4], [15, 12, 8], [7, 12, 4]]

------------------------------ Epoch 091 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.793230
- Iter 024 / 025, Loss: 0.891793

* Train accuracy / confusion: 57.88% / [[269, 61, 26], [105, 110, 53], [52, 40, 84]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [6, 11, 18], [6, 5, 12]]

------------------------------ Epoch 092 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.770794
- Iter 024 / 025, Loss: 1.082444

* Train accuracy / confusion: 59.75% / [[272, 63, 23], [94, 124, 48], [58, 36, 82]], 
* Val   accuracy / confusion: 52.88% / [[34, 6, 6], [11, 15, 9], [5, 12, 6]]

------------------------------ Epoch 093 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.904433
- Iter 024 / 025, Loss: 0.828029

* Train accuracy / confusion: 59.62% / [[264, 72, 18], [88, 133, 48], [63, 34, 80]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [11, 15, 9], [3, 12, 8]]

------------------------------ Epoch 094 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.841963
- Iter 024 / 025, Loss: 0.974642

* Train accuracy / confusion: 59.38% / [[273, 64, 20], [102, 125, 42], [54, 43, 77]], 
* Val   accuracy / confusion: 51.92% / [[36, 7, 3], [10, 12, 13], [9, 8, 6]]

------------------------------ Epoch 095 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.654049
- Iter 024 / 025, Loss: 0.757900

* Train accuracy / confusion: 62.88% / [[283, 57, 18], [81, 131, 52], [53, 36, 89]], 
* Val   accuracy / confusion: 50.96% / [[36, 8, 2], [16, 11, 8], [9, 8, 6]]

------------------------------ Epoch 096 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.871383
- Iter 024 / 025, Loss: 0.807324

* Train accuracy / confusion: 61.12% / [[274, 55, 30], [91, 124, 49], [46, 40, 91]], 
* Val   accuracy / confusion: 62.50% / [[35, 7, 4], [7, 20, 8], [3, 10, 10]]

------------------------------ Epoch 097 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.118600
- Iter 024 / 025, Loss: 0.653223

* Train accuracy / confusion: 62.38% / [[286, 45, 26], [100, 118, 50], [52, 28, 95]], 
* Val   accuracy / confusion: 48.08% / [[34, 9, 3], [14, 12, 9], [8, 11, 4]]

------------------------------ Epoch 098 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.856706
- Iter 024 / 025, Loss: 1.136675

* Train accuracy / confusion: 59.62% / [[272, 62, 24], [99, 117, 51], [46, 41, 88]], 
* Val   accuracy / confusion: 54.81% / [[34, 6, 6], [12, 13, 10], [3, 10, 10]]

------------------------------ Epoch 099 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.892962
- Iter 024 / 025, Loss: 0.919629

* Train accuracy / confusion: 59.25% / [[271, 67, 17], [93, 130, 48], [59, 42, 73]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [12, 17, 6], [3, 15, 5]]

------------------------------ Epoch 100 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.736152
- Iter 024 / 025, Loss: 0.945091

* Train accuracy / confusion: 60.88% / [[265, 72, 19], [85, 139, 46], [42, 49, 83]], 
* Val   accuracy / confusion: 50.00% / [[32, 9, 5], [11, 13, 11], [5, 11, 7]]

------------------------------ Epoch 101 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.741681
- Iter 024 / 025, Loss: 0.878898

* Train accuracy / confusion: 59.62% / [[270, 60, 31], [95, 118, 52], [50, 35, 89]], 
* Val   accuracy / confusion: 53.85% / [[32, 7, 7], [8, 17, 10], [6, 10, 7]]

------------------------------ Epoch 102 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.813410
- Iter 024 / 025, Loss: 0.808422

* Train accuracy / confusion: 60.38% / [[286, 48, 27], [98, 103, 61], [53, 30, 94]], 
* Val   accuracy / confusion: 56.73% / [[37, 6, 3], [11, 16, 8], [10, 7, 6]]

------------------------------ Epoch 103 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.802479
- Iter 024 / 025, Loss: 0.709858

* Train accuracy / confusion: 62.00% / [[283, 44, 29], [94, 116, 56], [55, 26, 97]], 
* Val   accuracy / confusion: 50.00% / [[30, 8, 8], [13, 13, 9], [4, 10, 9]]

------------------------------ Epoch 104 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.931677
- Iter 024 / 025, Loss: 0.792650

* Train accuracy / confusion: 61.75% / [[269, 67, 20], [89, 142, 37], [48, 45, 83]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [12, 17, 6], [7, 10, 6]]

------------------------------ Epoch 105 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.737431
- Iter 024 / 025, Loss: 0.695090

* Train accuracy / confusion: 61.12% / [[285, 54, 20], [104, 122, 45], [45, 43, 82]], 
* Val   accuracy / confusion: 49.04% / [[36, 7, 3], [13, 9, 13], [9, 8, 6]]

------------------------------ Epoch 106 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.878958
- Iter 024 / 025, Loss: 0.834414

* Train accuracy / confusion: 59.38% / [[258, 58, 38], [89, 130, 51], [52, 37, 87]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [7, 19, 9], [3, 14, 6]]

------------------------------ Epoch 107 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.904574
- Iter 024 / 025, Loss: 0.788513

* Train accuracy / confusion: 60.25% / [[261, 74, 22], [80, 137, 51], [49, 42, 84]], 
* Val   accuracy / confusion: 52.88% / [[29, 10, 7], [6, 16, 13], [2, 11, 10]]

------------------------------ Epoch 108 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.991119
- Iter 024 / 025, Loss: 0.795576

* Train accuracy / confusion: 62.38% / [[284, 49, 25], [91, 115, 58], [47, 31, 100]], 
* Val   accuracy / confusion: 52.88% / [[34, 6, 6], [11, 13, 11], [7, 8, 8]]

------------------------------ Epoch 109 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.812780
- Iter 024 / 025, Loss: 0.888312

* Train accuracy / confusion: 60.62% / [[276, 47, 32], [93, 115, 59], [46, 38, 94]], 
* Val   accuracy / confusion: 50.00% / [[33, 8, 5], [13, 13, 9], [6, 11, 6]]

------------------------------ Epoch 110 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.906292
- Iter 024 / 025, Loss: 0.785514

* Train accuracy / confusion: 62.25% / [[278, 49, 22], [89, 130, 55], [47, 40, 90]], 
* Val   accuracy / confusion: 53.85% / [[36, 7, 3], [13, 15, 7], [8, 10, 5]]

------------------------------ Epoch 111 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.824181
- Iter 024 / 025, Loss: 0.947350

* Train accuracy / confusion: 63.00% / [[279, 60, 18], [86, 135, 41], [48, 43, 90]], 
* Val   accuracy / confusion: 55.77% / [[34, 10, 2], [14, 16, 5], [6, 9, 8]]

------------------------------ Epoch 112 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.866245
- Iter 024 / 025, Loss: 0.825981

* Train accuracy / confusion: 63.00% / [[279, 57, 16], [86, 136, 50], [43, 44, 89]], 
* Val   accuracy / confusion: 50.00% / [[28, 15, 3], [12, 18, 5], [6, 11, 6]]

------------------------------ Epoch 113 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.960273
- Iter 024 / 025, Loss: 0.871304

* Train accuracy / confusion: 60.00% / [[275, 55, 22], [96, 121, 54], [56, 37, 84]], 
* Val   accuracy / confusion: 54.81% / [[36, 6, 4], [13, 13, 9], [8, 7, 8]]

------------------------------ Epoch 114 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.883485
- Iter 024 / 025, Loss: 0.816920

* Train accuracy / confusion: 61.38% / [[272, 65, 19], [89, 130, 50], [45, 41, 89]], 
* Val   accuracy / confusion: 53.85% / [[27, 14, 5], [6, 17, 12], [3, 8, 12]]

------------------------------ Epoch 115 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.910150
- Iter 024 / 025, Loss: 0.852636

* Train accuracy / confusion: 60.88% / [[283, 51, 21], [104, 115, 49], [48, 40, 89]], 
* Val   accuracy / confusion: 50.96% / [[36, 6, 4], [14, 10, 11], [8, 8, 7]]

------------------------------ Epoch 116 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.842090
- Iter 024 / 025, Loss: 0.846358

* Train accuracy / confusion: 63.88% / [[270, 58, 26], [79, 145, 44], [49, 33, 96]], 
* Val   accuracy / confusion: 50.00% / [[33, 10, 3], [12, 11, 12], [5, 10, 8]]

------------------------------ Epoch 117 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.621169
- Iter 024 / 025, Loss: 0.968750

* Train accuracy / confusion: 62.62% / [[272, 54, 27], [84, 153, 34], [46, 54, 76]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [10, 22, 3], [3, 16, 4]]

------------------------------ Epoch 118 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.734319
- Iter 024 / 025, Loss: 0.930030

* Train accuracy / confusion: 61.50% / [[269, 59, 27], [96, 124, 48], [44, 34, 99]], 
* Val   accuracy / confusion: 54.81% / [[36, 6, 4], [14, 14, 7], [7, 9, 7]]

------------------------------ Epoch 119 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.817762
- Iter 024 / 025, Loss: 0.869470

* Train accuracy / confusion: 63.12% / [[280, 54, 22], [93, 134, 40], [52, 34, 91]], 
* Val   accuracy / confusion: 56.73% / [[35, 7, 4], [7, 17, 11], [8, 8, 7]]

------------------------------ Epoch 120 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.681661
- Iter 024 / 025, Loss: 0.864909

* Train accuracy / confusion: 63.00% / [[274, 59, 22], [82, 135, 50], [44, 39, 95]], 
* Val   accuracy / confusion: 52.88% / [[34, 6, 6], [10, 15, 10], [9, 8, 6]]

------------------------------ Epoch 121 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.811552
- Iter 024 / 025, Loss: 0.863794

* Train accuracy / confusion: 64.75% / [[278, 46, 31], [84, 142, 45], [41, 35, 98]], 
* Val   accuracy / confusion: 49.04% / [[32, 8, 6], [12, 13, 10], [6, 11, 6]]

------------------------------ Epoch 122 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.957981
- Iter 024 / 025, Loss: 0.747622

* Train accuracy / confusion: 61.38% / [[285, 48, 23], [95, 125, 46], [52, 45, 81]], 
* Val   accuracy / confusion: 52.88% / [[27, 15, 4], [7, 23, 5], [3, 15, 5]]

------------------------------ Epoch 123 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.712251
- Iter 024 / 025, Loss: 0.704815

* Train accuracy / confusion: 62.00% / [[276, 61, 18], [85, 133, 52], [46, 42, 87]], 
* Val   accuracy / confusion: 56.73% / [[33, 12, 1], [11, 19, 5], [7, 9, 7]]

------------------------------ Epoch 124 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.969037
- Iter 024 / 025, Loss: 0.732393

* Train accuracy / confusion: 62.75% / [[290, 40, 25], [95, 128, 46], [57, 35, 84]], 
* Val   accuracy / confusion: 58.65% / [[34, 8, 4], [9, 16, 10], [4, 8, 11]]

------------------------------ Epoch 125 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.920323
- Iter 024 / 025, Loss: 0.874063

* Train accuracy / confusion: 63.12% / [[288, 51, 18], [91, 126, 48], [41, 46, 91]], 
* Val   accuracy / confusion: 61.54% / [[32, 11, 3], [5, 24, 6], [6, 9, 8]]

------------------------------ Epoch 126 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 1.016874
- Iter 024 / 025, Loss: 0.777857

* Train accuracy / confusion: 62.38% / [[264, 65, 24], [73, 153, 45], [47, 47, 82]], 
* Val   accuracy / confusion: 59.62% / [[34, 9, 3], [6, 19, 10], [4, 10, 9]]

------------------------------ Epoch 127 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.710738
- Iter 024 / 025, Loss: 0.745163

* Train accuracy / confusion: 62.88% / [[281, 45, 27], [90, 129, 50], [45, 40, 93]], 
* Val   accuracy / confusion: 54.81% / [[35, 7, 4], [9, 12, 14], [4, 9, 10]]

------------------------------ Epoch 128 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.776138
- Iter 024 / 025, Loss: 0.771229

* Train accuracy / confusion: 65.38% / [[286, 48, 21], [92, 131, 47], [45, 24, 106]], 
* Val   accuracy / confusion: 57.69% / [[33, 10, 3], [11, 18, 6], [6, 8, 9]]

------------------------------ Epoch 129 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.844772
- Iter 024 / 025, Loss: 0.779715

* Train accuracy / confusion: 68.00% / [[290, 55, 16], [70, 157, 36], [45, 34, 97]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [13, 16, 6], [8, 8, 7]]

------------------------------ Epoch 130 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.850784
- Iter 024 / 025, Loss: 0.749835

* Train accuracy / confusion: 61.50% / [[283, 47, 23], [90, 132, 49], [53, 46, 77]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [13, 17, 5], [7, 11, 5]]

------------------------------ Epoch 131 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.823987
- Iter 024 / 025, Loss: 0.822824

* Train accuracy / confusion: 64.12% / [[280, 46, 27], [84, 138, 52], [44, 34, 95]], 
* Val   accuracy / confusion: 57.69% / [[32, 11, 3], [11, 18, 6], [4, 9, 10]]

------------------------------ Epoch 132 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.953341
- Iter 024 / 025, Loss: 0.780973

* Train accuracy / confusion: 61.50% / [[285, 54, 19], [93, 131, 42], [58, 42, 76]], 
* Val   accuracy / confusion: 52.88% / [[32, 7, 7], [11, 15, 9], [9, 6, 8]]

------------------------------ Epoch 133 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.738045
- Iter 024 / 025, Loss: 0.653660

* Train accuracy / confusion: 64.50% / [[288, 40, 27], [92, 129, 48], [42, 35, 99]], 
* Val   accuracy / confusion: 53.85% / [[32, 8, 6], [10, 14, 11], [6, 7, 10]]

------------------------------ Epoch 134 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.799428
- Iter 024 / 025, Loss: 0.843907

* Train accuracy / confusion: 65.75% / [[290, 45, 22], [81, 141, 51], [39, 36, 95]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [12, 16, 7], [6, 8, 9]]

------------------------------ Epoch 135 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.827962
- Iter 024 / 025, Loss: 0.730458

* Train accuracy / confusion: 66.00% / [[294, 39, 22], [91, 148, 30], [52, 38, 86]], 
* Val   accuracy / confusion: 52.88% / [[33, 7, 6], [15, 11, 9], [8, 4, 11]]

------------------------------ Epoch 136 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.778602
- Iter 024 / 025, Loss: 0.918461

* Train accuracy / confusion: 65.50% / [[291, 38, 27], [88, 142, 40], [48, 35, 91]], 
* Val   accuracy / confusion: 47.12% / [[29, 13, 4], [11, 15, 9], [8, 10, 5]]

------------------------------ Epoch 137 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.864834
- Iter 024 / 025, Loss: 0.776892

* Train accuracy / confusion: 66.62% / [[296, 46, 19], [78, 150, 37], [53, 34, 87]], 
* Val   accuracy / confusion: 52.88% / [[34, 9, 3], [14, 14, 7], [9, 7, 7]]

------------------------------ Epoch 138 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.763126
- Iter 024 / 025, Loss: 0.688787

* Train accuracy / confusion: 66.50% / [[287, 51, 18], [77, 148, 42], [44, 36, 97]], 
* Val   accuracy / confusion: 57.69% / [[32, 9, 5], [8, 18, 9], [3, 10, 10]]

------------------------------ Epoch 139 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.684330
- Iter 024 / 025, Loss: 0.659846

* Train accuracy / confusion: 64.50% / [[283, 54, 21], [85, 134, 49], [47, 28, 99]], 
* Val   accuracy / confusion: 56.73% / [[37, 6, 3], [14, 16, 5], [6, 11, 6]]

------------------------------ Epoch 140 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.732920
- Iter 024 / 025, Loss: 0.674407

* Train accuracy / confusion: 64.88% / [[293, 39, 21], [99, 127, 44], [48, 30, 99]], 
* Val   accuracy / confusion: 47.12% / [[31, 9, 6], [17, 11, 7], [7, 9, 7]]

------------------------------ Epoch 141 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.796935
- Iter 024 / 025, Loss: 0.782001

* Train accuracy / confusion: 63.75% / [[288, 48, 24], [92, 132, 42], [46, 38, 90]], 
* Val   accuracy / confusion: 52.88% / [[36, 6, 4], [12, 12, 11], [6, 10, 7]]

------------------------------ Epoch 142 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.743710
- Iter 024 / 025, Loss: 0.749263

* Train accuracy / confusion: 67.25% / [[302, 39, 17], [90, 135, 43], [49, 24, 101]], 
* Val   accuracy / confusion: 56.73% / [[34, 10, 2], [12, 16, 7], [6, 8, 9]]

------------------------------ Epoch 143 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.805354
- Iter 024 / 025, Loss: 0.770737

* Train accuracy / confusion: 66.62% / [[293, 42, 25], [78, 143, 43], [40, 39, 97]], 
* Val   accuracy / confusion: 53.85% / [[36, 7, 3], [13, 17, 5], [10, 10, 3]]

------------------------------ Epoch 144 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.670168
- Iter 024 / 025, Loss: 0.862221

* Train accuracy / confusion: 66.00% / [[305, 34, 17], [90, 135, 41], [53, 37, 88]], 
* Val   accuracy / confusion: 57.69% / [[32, 10, 4], [11, 18, 6], [4, 9, 10]]

------------------------------ Epoch 145 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.533590
- Iter 024 / 025, Loss: 0.728993

* Train accuracy / confusion: 66.75% / [[287, 44, 19], [82, 139, 51], [42, 28, 108]], 
* Val   accuracy / confusion: 56.73% / [[35, 7, 4], [10, 15, 10], [5, 9, 9]]

------------------------------ Epoch 146 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.745293
- Iter 024 / 025, Loss: 0.988298

* Train accuracy / confusion: 66.62% / [[285, 44, 21], [88, 142, 42], [45, 27, 106]], 
* Val   accuracy / confusion: 57.69% / [[36, 7, 3], [13, 18, 4], [7, 10, 6]]

------------------------------ Epoch 147 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.695436
- Iter 024 / 025, Loss: 0.716003

* Train accuracy / confusion: 65.62% / [[284, 43, 29], [92, 143, 35], [44, 32, 98]], 
* Val   accuracy / confusion: 56.73% / [[33, 8, 5], [12, 20, 3], [7, 10, 6]]

------------------------------ Epoch 148 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.742069
- Iter 024 / 025, Loss: 0.687920

* Train accuracy / confusion: 65.88% / [[276, 46, 28], [75, 146, 50], [40, 34, 105]], 
* Val   accuracy / confusion: 55.77% / [[35, 8, 3], [11, 17, 7], [8, 9, 6]]

------------------------------ Epoch 149 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.837527
- Iter 024 / 025, Loss: 0.980899

* Train accuracy / confusion: 66.75% / [[288, 38, 28], [79, 142, 49], [39, 33, 104]], 
* Val   accuracy / confusion: 46.15% / [[28, 9, 9], [12, 7, 16], [3, 7, 13]]

------------------------------ Epoch 150 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.904669
- Iter 024 / 025, Loss: 0.697337

* Train accuracy / confusion: 66.75% / [[280, 53, 23], [73, 159, 37], [46, 34, 95]], 
* Val   accuracy / confusion: 53.85% / [[37, 6, 3], [16, 13, 6], [7, 10, 6]]

------------------------------ Epoch 151 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.959427
- Iter 024 / 025, Loss: 0.631877

* Train accuracy / confusion: 66.62% / [[302, 35, 18], [72, 151, 43], [46, 53, 80]], 
* Val   accuracy / confusion: 50.96% / [[36, 6, 4], [19, 12, 4], [9, 9, 5]]

------------------------------ Epoch 152 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.707292
- Iter 024 / 025, Loss: 0.821143

* Train accuracy / confusion: 67.88% / [[292, 49, 19], [71, 147, 47], [37, 34, 104]], 
* Val   accuracy / confusion: 53.85% / [[35, 6, 5], [16, 15, 4], [8, 9, 6]]

------------------------------ Epoch 153 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.702592
- Iter 024 / 025, Loss: 0.639964

* Train accuracy / confusion: 68.38% / [[302, 34, 21], [81, 144, 40], [43, 34, 101]], 
* Val   accuracy / confusion: 50.96% / [[32, 10, 4], [13, 13, 9], [9, 6, 8]]

------------------------------ Epoch 154 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.661500
- Iter 024 / 025, Loss: 0.673148

* Train accuracy / confusion: 69.12% / [[292, 40, 23], [80, 152, 38], [40, 26, 109]], 
* Val   accuracy / confusion: 53.85% / [[31, 8, 7], [12, 17, 6], [9, 6, 8]]

------------------------------ Epoch 155 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.700862
- Iter 024 / 025, Loss: 0.958302

* Train accuracy / confusion: 67.62% / [[294, 43, 20], [80, 155, 30], [46, 40, 92]], 
* Val   accuracy / confusion: 50.00% / [[32, 11, 3], [16, 9, 10], [6, 6, 11]]

------------------------------ Epoch 156 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.708428
- Iter 024 / 025, Loss: 0.564643

* Train accuracy / confusion: 71.25% / [[302, 33, 23], [74, 166, 32], [41, 27, 102]], 
* Val   accuracy / confusion: 55.77% / [[35, 9, 2], [13, 19, 3], [6, 13, 4]]

------------------------------ Epoch 157 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.619039
- Iter 024 / 025, Loss: 1.160918

* Train accuracy / confusion: 68.75% / [[297, 44, 17], [71, 160, 36], [43, 39, 93]], 
* Val   accuracy / confusion: 55.77% / [[34, 6, 6], [9, 17, 9], [8, 8, 7]]

------------------------------ Epoch 158 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.836050
- Iter 024 / 025, Loss: 0.802797

* Train accuracy / confusion: 66.00% / [[290, 44, 22], [87, 143, 38], [42, 39, 95]], 
* Val   accuracy / confusion: 50.96% / [[34, 10, 2], [10, 14, 11], [11, 7, 5]]

------------------------------ Epoch 159 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.719372
- Iter 024 / 025, Loss: 0.535852

* Train accuracy / confusion: 68.25% / [[291, 48, 24], [72, 154, 39], [42, 29, 101]], 
* Val   accuracy / confusion: 54.81% / [[27, 16, 3], [4, 23, 8], [6, 10, 7]]

------------------------------ Epoch 160 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.796074
- Iter 024 / 025, Loss: 0.693795

* Train accuracy / confusion: 70.38% / [[306, 33, 18], [70, 164, 34], [47, 35, 93]], 
* Val   accuracy / confusion: 59.62% / [[38, 6, 2], [13, 19, 3], [9, 9, 5]]

------------------------------ Epoch 161 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.554363
- Iter 024 / 025, Loss: 0.760766

* Train accuracy / confusion: 70.00% / [[296, 41, 17], [71, 162, 36], [39, 36, 102]], 
* Val   accuracy / confusion: 50.00% / [[29, 11, 6], [10, 14, 11], [5, 9, 9]]

------------------------------ Epoch 162 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.533427
- Iter 024 / 025, Loss: 0.887156

* Train accuracy / confusion: 68.88% / [[293, 46, 18], [74, 161, 38], [51, 22, 97]], 
* Val   accuracy / confusion: 54.81% / [[31, 10, 5], [8, 18, 9], [1, 14, 8]]

------------------------------ Epoch 163 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.601873
- Iter 024 / 025, Loss: 0.769605

* Train accuracy / confusion: 68.38% / [[284, 41, 29], [73, 165, 29], [30, 51, 98]], 
* Val   accuracy / confusion: 55.77% / [[33, 8, 5], [8, 19, 8], [5, 12, 6]]

------------------------------ Epoch 164 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.868561
- Iter 024 / 025, Loss: 1.065211

* Train accuracy / confusion: 68.25% / [[303, 28, 25], [80, 141, 45], [42, 34, 102]], 
* Val   accuracy / confusion: 55.77% / [[33, 7, 6], [10, 19, 6], [3, 14, 6]]

------------------------------ Epoch 165 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.592111
- Iter 024 / 025, Loss: 0.618585

* Train accuracy / confusion: 68.00% / [[286, 47, 20], [76, 156, 37], [46, 30, 102]], 
* Val   accuracy / confusion: 55.77% / [[31, 12, 3], [6, 19, 10], [7, 8, 8]]

------------------------------ Epoch 166 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.505678
- Iter 024 / 025, Loss: 0.699065

* Train accuracy / confusion: 69.00% / [[292, 43, 18], [78, 160, 34], [36, 39, 100]], 
* Val   accuracy / confusion: 50.96% / [[32, 10, 4], [14, 15, 6], [8, 9, 6]]

------------------------------ Epoch 167 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.740748
- Iter 024 / 025, Loss: 0.659949

* Train accuracy / confusion: 68.75% / [[292, 39, 24], [70, 162, 36], [35, 46, 96]], 
* Val   accuracy / confusion: 58.65% / [[34, 9, 3], [11, 19, 5], [7, 8, 8]]

------------------------------ Epoch 168 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.635951
- Iter 024 / 025, Loss: 0.838221

* Train accuracy / confusion: 67.75% / [[306, 34, 19], [92, 135, 39], [42, 32, 101]], 
* Val   accuracy / confusion: 51.92% / [[32, 8, 6], [9, 12, 14], [5, 8, 10]]

------------------------------ Epoch 169 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.548648
- Iter 024 / 025, Loss: 0.592852

* Train accuracy / confusion: 69.62% / [[294, 43, 17], [78, 155, 38], [37, 30, 108]], 
* Val   accuracy / confusion: 54.81% / [[29, 11, 6], [10, 19, 6], [4, 10, 9]]

------------------------------ Epoch 170 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.694021
- Iter 024 / 025, Loss: 0.632335

* Train accuracy / confusion: 69.25% / [[285, 40, 30], [70, 166, 32], [37, 37, 103]], 
* Val   accuracy / confusion: 58.65% / [[32, 10, 4], [8, 18, 9], [2, 10, 11]]

------------------------------ Epoch 171 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.574835
- Iter 024 / 025, Loss: 0.720785

* Train accuracy / confusion: 71.62% / [[292, 49, 18], [70, 162, 35], [30, 25, 119]], 
* Val   accuracy / confusion: 58.65% / [[34, 8, 4], [8, 17, 10], [2, 11, 10]]

------------------------------ Epoch 172 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.656497
- Iter 024 / 025, Loss: 0.770228

* Train accuracy / confusion: 67.50% / [[269, 50, 34], [75, 161, 35], [35, 31, 110]], 
* Val   accuracy / confusion: 58.65% / [[30, 9, 7], [5, 20, 10], [1, 11, 11]]

------------------------------ Epoch 173 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.808817
- Iter 024 / 025, Loss: 0.749771

* Train accuracy / confusion: 69.50% / [[302, 30, 26], [86, 138, 40], [33, 29, 116]], 
* Val   accuracy / confusion: 55.77% / [[37, 5, 4], [19, 13, 3], [14, 1, 8]]

------------------------------ Epoch 174 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.786630
- Iter 024 / 025, Loss: 0.670581

* Train accuracy / confusion: 71.62% / [[291, 46, 20], [64, 173, 29], [32, 36, 109]], 
* Val   accuracy / confusion: 48.08% / [[32, 10, 4], [18, 14, 3], [9, 10, 4]]

------------------------------ Epoch 175 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.685353
- Iter 024 / 025, Loss: 0.516708

* Train accuracy / confusion: 72.62% / [[294, 42, 19], [62, 169, 39], [30, 27, 118]], 
* Val   accuracy / confusion: 53.85% / [[34, 7, 5], [17, 12, 6], [4, 9, 10]]

------------------------------ Epoch 176 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.756600
- Iter 024 / 025, Loss: 0.533462

* Train accuracy / confusion: 70.12% / [[297, 39, 20], [82, 152, 33], [33, 32, 112]], 
* Val   accuracy / confusion: 55.77% / [[30, 11, 5], [12, 15, 8], [4, 6, 13]]

------------------------------ Epoch 177 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.650032
- Iter 024 / 025, Loss: 0.572900

* Train accuracy / confusion: 71.62% / [[294, 44, 18], [64, 175, 31], [35, 35, 104]], 
* Val   accuracy / confusion: 50.96% / [[36, 6, 4], [13, 12, 10], [8, 10, 5]]

------------------------------ Epoch 178 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.671006
- Iter 024 / 025, Loss: 0.620096

* Train accuracy / confusion: 72.62% / [[305, 35, 15], [63, 167, 38], [38, 30, 109]], 
* Val   accuracy / confusion: 53.85% / [[30, 7, 9], [8, 17, 10], [2, 12, 9]]

------------------------------ Epoch 179 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.615990
- Iter 024 / 025, Loss: 0.776390

* Train accuracy / confusion: 72.00% / [[288, 45, 24], [78, 166, 21], [23, 33, 122]], 
* Val   accuracy / confusion: 50.00% / [[32, 8, 6], [11, 12, 12], [8, 7, 8]]

------------------------------ Epoch 180 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.570707
- Iter 024 / 025, Loss: 0.786395

* Train accuracy / confusion: 75.00% / [[309, 33, 16], [71, 170, 24], [33, 23, 121]], 
* Val   accuracy / confusion: 50.00% / [[34, 7, 5], [16, 13, 6], [9, 9, 5]]

------------------------------ Epoch 181 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.992449
- Iter 024 / 025, Loss: 0.525064

* Train accuracy / confusion: 68.62% / [[276, 61, 23], [64, 171, 33], [41, 29, 102]], 
* Val   accuracy / confusion: 59.62% / [[32, 9, 5], [9, 19, 7], [6, 6, 11]]

------------------------------ Epoch 182 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.707562
- Iter 024 / 025, Loss: 0.802462

* Train accuracy / confusion: 71.50% / [[294, 39, 18], [68, 167, 35], [31, 37, 111]], 
* Val   accuracy / confusion: 53.85% / [[30, 9, 7], [13, 17, 5], [7, 7, 9]]

------------------------------ Epoch 183 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.639033
- Iter 024 / 025, Loss: 0.550069

* Train accuracy / confusion: 74.75% / [[305, 34, 20], [58, 184, 27], [26, 37, 109]], 
* Val   accuracy / confusion: 55.77% / [[29, 11, 6], [6, 19, 10], [3, 10, 10]]

------------------------------ Epoch 184 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.631667
- Iter 024 / 025, Loss: 0.566339

* Train accuracy / confusion: 72.12% / [[295, 45, 18], [68, 165, 34], [28, 30, 117]], 
* Val   accuracy / confusion: 52.88% / [[30, 8, 8], [9, 15, 11], [3, 10, 10]]

------------------------------ Epoch 185 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.716665
- Iter 024 / 025, Loss: 0.506612

* Train accuracy / confusion: 70.12% / [[304, 36, 19], [84, 149, 31], [31, 38, 108]], 
* Val   accuracy / confusion: 59.62% / [[30, 9, 7], [8, 20, 7], [6, 5, 12]]

------------------------------ Epoch 186 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.617171
- Iter 024 / 025, Loss: 0.962789

* Train accuracy / confusion: 72.00% / [[293, 37, 24], [61, 167, 39], [28, 35, 116]], 
* Val   accuracy / confusion: 53.85% / [[34, 6, 6], [12, 15, 8], [3, 13, 7]]

------------------------------ Epoch 187 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.574543
- Iter 024 / 025, Loss: 0.748772

* Train accuracy / confusion: 72.50% / [[291, 42, 22], [59, 181, 30], [33, 34, 108]], 
* Val   accuracy / confusion: 50.96% / [[31, 10, 5], [9, 15, 11], [4, 12, 7]]

------------------------------ Epoch 188 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.602764
- Iter 024 / 025, Loss: 0.748538

* Train accuracy / confusion: 71.75% / [[296, 40, 20], [58, 170, 40], [31, 37, 108]], 
* Val   accuracy / confusion: 47.12% / [[30, 10, 6], [17, 14, 4], [7, 11, 5]]

------------------------------ Epoch 189 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.753395
- Iter 024 / 025, Loss: 0.853857

* Train accuracy / confusion: 73.12% / [[291, 37, 24], [60, 182, 29], [35, 30, 112]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [9, 15, 11], [9, 7, 7]]

------------------------------ Epoch 190 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.650069
- Iter 024 / 025, Loss: 0.461406

* Train accuracy / confusion: 73.88% / [[299, 44, 13], [57, 177, 33], [28, 34, 115]], 
* Val   accuracy / confusion: 50.00% / [[30, 11, 5], [13, 13, 9], [6, 8, 9]]

------------------------------ Epoch 191 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.770812
- Iter 024 / 025, Loss: 0.719751

* Train accuracy / confusion: 73.12% / [[297, 38, 17], [64, 173, 35], [31, 30, 115]], 
* Val   accuracy / confusion: 49.04% / [[31, 12, 3], [14, 14, 7], [6, 11, 6]]

------------------------------ Epoch 192 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.537557
- Iter 024 / 025, Loss: 0.669184

* Train accuracy / confusion: 73.62% / [[303, 38, 15], [59, 178, 33], [31, 35, 108]], 
* Val   accuracy / confusion: 50.00% / [[25, 11, 10], [8, 16, 11], [4, 8, 11]]

------------------------------ Epoch 193 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.599026
- Iter 024 / 025, Loss: 0.612630

* Train accuracy / confusion: 74.25% / [[299, 40, 13], [70, 179, 22], [33, 28, 116]], 
* Val   accuracy / confusion: 55.77% / [[36, 4, 6], [10, 16, 9], [10, 7, 6]]

------------------------------ Epoch 194 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.937026
- Iter 024 / 025, Loss: 0.760276

* Train accuracy / confusion: 72.12% / [[293, 45, 19], [58, 176, 33], [39, 29, 108]], 
* Val   accuracy / confusion: 54.81% / [[29, 10, 7], [8, 18, 9], [7, 6, 10]]

------------------------------ Epoch 195 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.765920
- Iter 024 / 025, Loss: 0.498812

* Train accuracy / confusion: 74.88% / [[302, 34, 18], [63, 186, 22], [29, 35, 111]], 
* Val   accuracy / confusion: 55.77% / [[31, 10, 5], [10, 17, 8], [5, 8, 10]]

------------------------------ Epoch 196 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.975353
- Iter 024 / 025, Loss: 0.636155

* Train accuracy / confusion: 73.50% / [[307, 33, 18], [75, 169, 24], [31, 31, 112]], 
* Val   accuracy / confusion: 52.88% / [[31, 10, 5], [14, 17, 4], [3, 13, 7]]

------------------------------ Epoch 197 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.743747
- Iter 024 / 025, Loss: 0.486287

* Train accuracy / confusion: 73.88% / [[298, 44, 15], [51, 179, 38], [35, 26, 114]], 
* Val   accuracy / confusion: 49.04% / [[28, 10, 8], [10, 13, 12], [6, 7, 10]]

------------------------------ Epoch 198 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.567228
- Iter 024 / 025, Loss: 0.924077

* Train accuracy / confusion: 73.38% / [[300, 39, 16], [59, 178, 30], [28, 41, 109]], 
* Val   accuracy / confusion: 53.85% / [[30, 9, 7], [10, 19, 6], [6, 10, 7]]

------------------------------ Epoch 199 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.594772
- Iter 024 / 025, Loss: 0.532770

* Train accuracy / confusion: 76.50% / [[295, 37, 21], [53, 183, 32], [28, 17, 134]], 
* Val   accuracy / confusion: 58.65% / [[34, 8, 4], [10, 19, 6], [6, 9, 8]]

------------------------------ Epoch 200 / 500, Learning rate: 3.81e-03 ------------------------------
- Iter 012 / 025, Loss: 0.800291
- Iter 024 / 025, Loss: 0.310671

* Train accuracy / confusion: 76.38% / [[306, 31, 16], [63, 176, 28], [20, 31, 129]], 
* Val   accuracy / confusion: 52.88% / [[31, 9, 6], [6, 14, 15], [4, 9, 10]]

------------------------------ Epoch 201 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.577357
- Iter 024 / 025, Loss: 0.516436

* Train accuracy / confusion: 76.88% / [[305, 37, 15], [58, 179, 28], [30, 17, 131]], 
* Val   accuracy / confusion: 60.58% / [[31, 8, 7], [11, 16, 8], [2, 5, 16]]

------------------------------ Epoch 202 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.511414
- Iter 024 / 025, Loss: 0.567988

* Train accuracy / confusion: 76.12% / [[322, 25, 12], [70, 165, 31], [29, 24, 122]], 
* Val   accuracy / confusion: 50.96% / [[30, 10, 6], [10, 16, 9], [9, 7, 7]]

------------------------------ Epoch 203 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.661900
- Iter 024 / 025, Loss: 0.377975

* Train accuracy / confusion: 79.75% / [[320, 25, 11], [55, 188, 23], [32, 16, 130]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [9, 16, 10], [5, 7, 11]]

------------------------------ Epoch 204 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.591135
- Iter 024 / 025, Loss: 0.466667

* Train accuracy / confusion: 79.12% / [[323, 25, 8], [56, 187, 27], [19, 32, 123]], 
* Val   accuracy / confusion: 53.85% / [[30, 8, 8], [8, 18, 9], [7, 8, 8]]

------------------------------ Epoch 205 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.416805
- Iter 024 / 025, Loss: 0.693693

* Train accuracy / confusion: 78.62% / [[319, 32, 7], [49, 191, 31], [32, 20, 119]], 
* Val   accuracy / confusion: 53.85% / [[33, 7, 6], [14, 14, 7], [8, 6, 9]]

------------------------------ Epoch 206 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.389832
- Iter 024 / 025, Loss: 0.484955

* Train accuracy / confusion: 76.88% / [[305, 33, 15], [51, 189, 31], [25, 30, 121]], 
* Val   accuracy / confusion: 61.54% / [[35, 6, 5], [10, 18, 7], [6, 6, 11]]

------------------------------ Epoch 207 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.429427
- Iter 024 / 025, Loss: 0.412180

* Train accuracy / confusion: 77.38% / [[306, 36, 12], [53, 184, 30], [25, 25, 129]], 
* Val   accuracy / confusion: 46.15% / [[27, 14, 5], [11, 16, 8], [8, 10, 5]]

------------------------------ Epoch 208 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.393720
- Iter 024 / 025, Loss: 0.695525

* Train accuracy / confusion: 78.12% / [[309, 32, 10], [54, 187, 32], [19, 28, 129]], 
* Val   accuracy / confusion: 55.77% / [[30, 13, 3], [12, 18, 5], [5, 8, 10]]

------------------------------ Epoch 209 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.399691
- Iter 024 / 025, Loss: 0.372400

* Train accuracy / confusion: 81.25% / [[316, 28, 11], [50, 203, 18], [21, 22, 131]], 
* Val   accuracy / confusion: 52.88% / [[33, 7, 6], [9, 15, 11], [9, 7, 7]]

------------------------------ Epoch 210 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.555527
- Iter 024 / 025, Loss: 0.753473

* Train accuracy / confusion: 78.88% / [[308, 36, 9], [54, 197, 23], [23, 24, 126]], 
* Val   accuracy / confusion: 52.88% / [[31, 9, 6], [14, 16, 5], [7, 8, 8]]

------------------------------ Epoch 211 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.514313
- Iter 024 / 025, Loss: 0.461418

* Train accuracy / confusion: 78.00% / [[311, 30, 12], [49, 192, 32], [24, 29, 121]], 
* Val   accuracy / confusion: 58.65% / [[33, 9, 4], [10, 18, 7], [7, 6, 10]]

------------------------------ Epoch 212 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.676120
- Iter 024 / 025, Loss: 0.574294

* Train accuracy / confusion: 78.75% / [[315, 34, 11], [53, 186, 25], [21, 26, 129]], 
* Val   accuracy / confusion: 50.00% / [[30, 12, 4], [13, 17, 5], [9, 9, 5]]

------------------------------ Epoch 213 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.648043
- Iter 024 / 025, Loss: 0.499733

* Train accuracy / confusion: 77.38% / [[307, 33, 10], [60, 184, 27], [26, 25, 128]], 
* Val   accuracy / confusion: 49.04% / [[32, 8, 6], [19, 13, 3], [10, 7, 6]]

------------------------------ Epoch 214 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.476428
- Iter 024 / 025, Loss: 0.549603

* Train accuracy / confusion: 78.25% / [[312, 43, 7], [54, 181, 24], [23, 23, 133]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 16, 12], [8, 7, 8]]

------------------------------ Epoch 215 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.387125
- Iter 024 / 025, Loss: 0.415990

* Train accuracy / confusion: 80.12% / [[315, 33, 9], [49, 193, 21], [22, 25, 133]], 
* Val   accuracy / confusion: 52.88% / [[29, 11, 6], [9, 16, 10], [4, 9, 10]]

------------------------------ Epoch 216 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.521659
- Iter 024 / 025, Loss: 0.553212

* Train accuracy / confusion: 80.12% / [[316, 28, 13], [43, 195, 29], [20, 26, 130]], 
* Val   accuracy / confusion: 51.92% / [[30, 12, 4], [9, 15, 11], [9, 5, 9]]

------------------------------ Epoch 217 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.612657
- Iter 024 / 025, Loss: 0.629088

* Train accuracy / confusion: 80.62% / [[319, 23, 16], [43, 197, 26], [22, 25, 129]], 
* Val   accuracy / confusion: 50.96% / [[31, 11, 4], [17, 12, 6], [9, 4, 10]]

------------------------------ Epoch 218 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.418519
- Iter 024 / 025, Loss: 0.435876

* Train accuracy / confusion: 80.62% / [[311, 32, 14], [41, 204, 23], [21, 24, 130]], 
* Val   accuracy / confusion: 64.42% / [[34, 7, 5], [7, 22, 6], [5, 7, 11]]

------------------------------ Epoch 219 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.468421
- Iter 024 / 025, Loss: 0.459749

* Train accuracy / confusion: 80.12% / [[314, 29, 12], [49, 193, 28], [19, 22, 134]], 
* Val   accuracy / confusion: 50.96% / [[29, 10, 7], [11, 14, 10], [8, 5, 10]]

------------------------------ Epoch 220 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.376749
- Iter 024 / 025, Loss: 0.325085

* Train accuracy / confusion: 81.00% / [[313, 26, 17], [38, 205, 26], [22, 23, 130]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [13, 15, 7], [6, 7, 10]]

------------------------------ Epoch 221 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.612649
- Iter 024 / 025, Loss: 0.520186

* Train accuracy / confusion: 81.62% / [[315, 28, 12], [47, 201, 20], [20, 20, 137]], 
* Val   accuracy / confusion: 51.92% / [[31, 8, 7], [14, 16, 5], [7, 9, 7]]

------------------------------ Epoch 222 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.399026
- Iter 024 / 025, Loss: 0.273557

* Train accuracy / confusion: 81.75% / [[314, 34, 11], [42, 207, 19], [15, 25, 133]], 
* Val   accuracy / confusion: 58.65% / [[32, 9, 5], [11, 20, 4], [6, 8, 9]]

------------------------------ Epoch 223 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.591615
- Iter 024 / 025, Loss: 0.394440

* Train accuracy / confusion: 82.38% / [[320, 27, 8], [52, 198, 17], [18, 19, 141]], 
* Val   accuracy / confusion: 49.04% / [[30, 11, 5], [14, 13, 8], [11, 4, 8]]

------------------------------ Epoch 224 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.366778
- Iter 024 / 025, Loss: 0.445582

* Train accuracy / confusion: 81.00% / [[312, 33, 12], [48, 203, 16], [22, 21, 133]], 
* Val   accuracy / confusion: 53.85% / [[30, 11, 5], [10, 19, 6], [10, 6, 7]]

------------------------------ Epoch 225 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.566488
- Iter 024 / 025, Loss: 0.248616

* Train accuracy / confusion: 80.12% / [[312, 31, 13], [46, 195, 25], [17, 27, 134]], 
* Val   accuracy / confusion: 55.77% / [[29, 11, 6], [8, 19, 8], [5, 8, 10]]

------------------------------ Epoch 226 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.469908
- Iter 024 / 025, Loss: 0.476905

* Train accuracy / confusion: 81.25% / [[317, 30, 9], [46, 198, 24], [25, 16, 135]], 
* Val   accuracy / confusion: 53.85% / [[32, 9, 5], [9, 18, 8], [5, 12, 6]]

------------------------------ Epoch 227 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.621254
- Iter 024 / 025, Loss: 0.550585

* Train accuracy / confusion: 81.50% / [[319, 31, 10], [42, 200, 24], [21, 20, 133]], 
* Val   accuracy / confusion: 50.96% / [[29, 13, 4], [11, 15, 9], [4, 10, 9]]

------------------------------ Epoch 228 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.525584
- Iter 024 / 025, Loss: 0.443244

* Train accuracy / confusion: 81.38% / [[314, 32, 15], [42, 202, 21], [17, 22, 135]], 
* Val   accuracy / confusion: 56.73% / [[32, 10, 4], [11, 16, 8], [7, 5, 11]]

------------------------------ Epoch 229 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.570001
- Iter 024 / 025, Loss: 0.419788

* Train accuracy / confusion: 79.12% / [[312, 23, 19], [44, 192, 30], [25, 26, 129]], 
* Val   accuracy / confusion: 58.65% / [[36, 5, 5], [11, 17, 7], [7, 8, 8]]

------------------------------ Epoch 230 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.422837
- Iter 024 / 025, Loss: 0.437616

* Train accuracy / confusion: 80.75% / [[317, 29, 11], [42, 201, 24], [18, 30, 128]], 
* Val   accuracy / confusion: 52.88% / [[33, 7, 6], [14, 15, 6], [10, 6, 7]]

------------------------------ Epoch 231 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.376724
- Iter 024 / 025, Loss: 0.930089

* Train accuracy / confusion: 81.25% / [[305, 34, 12], [41, 207, 24], [20, 19, 138]], 
* Val   accuracy / confusion: 56.73% / [[29, 13, 4], [11, 17, 7], [6, 4, 13]]

------------------------------ Epoch 232 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.421917
- Iter 024 / 025, Loss: 0.340115

* Train accuracy / confusion: 80.62% / [[312, 26, 12], [36, 212, 26], [31, 24, 121]], 
* Val   accuracy / confusion: 57.69% / [[32, 7, 7], [10, 16, 9], [7, 4, 12]]

------------------------------ Epoch 233 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.491746
- Iter 024 / 025, Loss: 0.216251

* Train accuracy / confusion: 81.62% / [[312, 31, 10], [42, 205, 23], [22, 19, 136]], 
* Val   accuracy / confusion: 50.96% / [[28, 12, 6], [11, 15, 9], [6, 7, 10]]

------------------------------ Epoch 234 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.483825
- Iter 024 / 025, Loss: 0.449399

* Train accuracy / confusion: 81.50% / [[321, 25, 12], [43, 201, 26], [16, 26, 130]], 
* Val   accuracy / confusion: 52.88% / [[32, 9, 5], [16, 15, 4], [8, 7, 8]]

------------------------------ Epoch 235 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.308565
- Iter 024 / 025, Loss: 0.389244

* Train accuracy / confusion: 82.50% / [[317, 30, 8], [37, 210, 24], [17, 24, 133]], 
* Val   accuracy / confusion: 51.92% / [[32, 8, 6], [10, 14, 11], [7, 8, 8]]

------------------------------ Epoch 236 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.513887
- Iter 024 / 025, Loss: 0.330703

* Train accuracy / confusion: 83.00% / [[320, 25, 12], [44, 207, 18], [15, 22, 137]], 
* Val   accuracy / confusion: 52.88% / [[29, 13, 4], [11, 18, 6], [6, 9, 8]]

------------------------------ Epoch 237 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.371242
- Iter 024 / 025, Loss: 0.502057

* Train accuracy / confusion: 80.25% / [[303, 34, 16], [45, 197, 24], [19, 20, 142]], 
* Val   accuracy / confusion: 58.65% / [[33, 8, 5], [14, 17, 4], [5, 7, 11]]

------------------------------ Epoch 238 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.697490
- Iter 024 / 025, Loss: 0.391045

* Train accuracy / confusion: 82.50% / [[318, 31, 7], [39, 210, 21], [19, 23, 132]], 
* Val   accuracy / confusion: 46.15% / [[28, 11, 7], [14, 13, 8], [6, 10, 7]]

------------------------------ Epoch 239 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.348404
- Iter 024 / 025, Loss: 0.347141

* Train accuracy / confusion: 80.62% / [[313, 30, 13], [45, 199, 25], [16, 26, 133]], 
* Val   accuracy / confusion: 50.96% / [[30, 8, 8], [15, 15, 5], [7, 8, 8]]

------------------------------ Epoch 240 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.233694
- Iter 024 / 025, Loss: 0.549187

* Train accuracy / confusion: 81.75% / [[317, 28, 14], [38, 203, 25], [20, 21, 134]], 
* Val   accuracy / confusion: 56.73% / [[32, 9, 5], [9, 19, 7], [8, 7, 8]]

------------------------------ Epoch 241 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.691021
- Iter 024 / 025, Loss: 0.352943

* Train accuracy / confusion: 82.62% / [[312, 29, 15], [36, 212, 20], [23, 16, 137]], 
* Val   accuracy / confusion: 53.85% / [[30, 12, 4], [11, 18, 6], [7, 8, 8]]

------------------------------ Epoch 242 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.305256
- Iter 024 / 025, Loss: 0.360522

* Train accuracy / confusion: 83.25% / [[318, 21, 14], [38, 213, 17], [21, 23, 135]], 
* Val   accuracy / confusion: 52.88% / [[26, 14, 6], [9, 18, 8], [6, 6, 11]]

------------------------------ Epoch 243 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.447292
- Iter 024 / 025, Loss: 0.593229

* Train accuracy / confusion: 82.12% / [[308, 33, 12], [44, 209, 16], [21, 17, 140]], 
* Val   accuracy / confusion: 55.77% / [[30, 10, 6], [10, 17, 8], [7, 5, 11]]

------------------------------ Epoch 244 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.336981
- Iter 024 / 025, Loss: 0.514027

* Train accuracy / confusion: 81.12% / [[313, 29, 14], [44, 201, 24], [19, 21, 135]], 
* Val   accuracy / confusion: 54.81% / [[28, 12, 6], [9, 17, 9], [5, 6, 12]]

------------------------------ Epoch 245 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.640602
- Iter 024 / 025, Loss: 0.529213

* Train accuracy / confusion: 83.00% / [[311, 31, 10], [38, 212, 20], [16, 21, 141]], 
* Val   accuracy / confusion: 50.00% / [[31, 11, 4], [15, 15, 5], [9, 8, 6]]

------------------------------ Epoch 246 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.607784
- Iter 024 / 025, Loss: 0.524207

* Train accuracy / confusion: 82.00% / [[325, 24, 9], [34, 201, 31], [18, 28, 130]], 
* Val   accuracy / confusion: 53.85% / [[29, 10, 7], [12, 16, 7], [1, 11, 11]]

------------------------------ Epoch 247 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.475364
- Iter 024 / 025, Loss: 0.432908

* Train accuracy / confusion: 81.12% / [[311, 30, 15], [45, 200, 23], [17, 21, 138]], 
* Val   accuracy / confusion: 50.96% / [[33, 8, 5], [16, 12, 7], [7, 8, 8]]

------------------------------ Epoch 248 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.374588
- Iter 024 / 025, Loss: 0.531906

* Train accuracy / confusion: 81.75% / [[313, 31, 10], [49, 204, 17], [17, 22, 137]], 
* Val   accuracy / confusion: 57.69% / [[33, 6, 7], [10, 15, 10], [4, 7, 12]]

------------------------------ Epoch 249 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.513520
- Iter 024 / 025, Loss: 0.383028

* Train accuracy / confusion: 82.62% / [[320, 23, 15], [42, 204, 24], [15, 20, 137]], 
* Val   accuracy / confusion: 56.73% / [[33, 6, 7], [9, 17, 9], [7, 7, 9]]

------------------------------ Epoch 250 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.239745
- Iter 024 / 025, Loss: 0.832758

* Train accuracy / confusion: 80.38% / [[316, 31, 10], [46, 197, 25], [25, 20, 130]], 
* Val   accuracy / confusion: 60.58% / [[34, 9, 3], [8, 19, 8], [4, 9, 10]]

------------------------------ Epoch 251 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.279443
- Iter 024 / 025, Loss: 0.396490

* Train accuracy / confusion: 83.88% / [[330, 18, 11], [46, 204, 20], [13, 21, 137]], 
* Val   accuracy / confusion: 57.69% / [[31, 11, 4], [10, 18, 7], [2, 10, 11]]

------------------------------ Epoch 252 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.396759
- Iter 024 / 025, Loss: 0.364778

* Train accuracy / confusion: 82.25% / [[313, 32, 11], [39, 212, 18], [19, 23, 133]], 
* Val   accuracy / confusion: 52.88% / [[27, 10, 9], [8, 16, 11], [4, 7, 12]]

------------------------------ Epoch 253 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.379693
- Iter 024 / 025, Loss: 0.301422

* Train accuracy / confusion: 80.00% / [[306, 32, 15], [46, 198, 23], [22, 22, 136]], 
* Val   accuracy / confusion: 57.69% / [[30, 11, 5], [9, 19, 7], [5, 7, 11]]

------------------------------ Epoch 254 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.254034
- Iter 024 / 025, Loss: 0.496701

* Train accuracy / confusion: 82.88% / [[323, 22, 9], [37, 210, 24], [19, 26, 130]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [13, 16, 6], [8, 6, 9]]

------------------------------ Epoch 255 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.375935
- Iter 024 / 025, Loss: 0.286166

* Train accuracy / confusion: 82.38% / [[319, 30, 7], [45, 205, 19], [15, 25, 135]], 
* Val   accuracy / confusion: 59.62% / [[27, 13, 6], [10, 19, 6], [4, 3, 16]]

------------------------------ Epoch 256 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.301433
- Iter 024 / 025, Loss: 0.402291

* Train accuracy / confusion: 82.75% / [[316, 29, 9], [41, 208, 20], [21, 18, 138]], 
* Val   accuracy / confusion: 56.73% / [[28, 10, 8], [11, 17, 7], [6, 3, 14]]

------------------------------ Epoch 257 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.453749
- Iter 024 / 025, Loss: 0.358465

* Train accuracy / confusion: 81.88% / [[315, 34, 8], [49, 201, 20], [12, 22, 139]], 
* Val   accuracy / confusion: 50.96% / [[31, 11, 4], [13, 14, 8], [9, 6, 8]]

------------------------------ Epoch 258 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.441417
- Iter 024 / 025, Loss: 0.483658

* Train accuracy / confusion: 83.25% / [[323, 25, 9], [42, 206, 19], [13, 26, 137]], 
* Val   accuracy / confusion: 60.58% / [[30, 14, 2], [6, 22, 7], [7, 5, 11]]

------------------------------ Epoch 259 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.324663
- Iter 024 / 025, Loss: 0.497563

* Train accuracy / confusion: 82.25% / [[321, 23, 12], [50, 199, 18], [20, 19, 138]], 
* Val   accuracy / confusion: 48.08% / [[27, 9, 10], [8, 14, 13], [7, 7, 9]]

------------------------------ Epoch 260 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.334226
- Iter 024 / 025, Loss: 0.520526

* Train accuracy / confusion: 83.00% / [[320, 22, 13], [42, 206, 19], [20, 20, 138]], 
* Val   accuracy / confusion: 50.96% / [[28, 11, 7], [11, 15, 9], [7, 6, 10]]

------------------------------ Epoch 261 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.530691
- Iter 024 / 025, Loss: 0.424063

* Train accuracy / confusion: 82.00% / [[314, 30, 8], [43, 201, 24], [14, 25, 141]], 
* Val   accuracy / confusion: 57.69% / [[29, 12, 5], [7, 20, 8], [5, 7, 11]]

------------------------------ Epoch 262 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.609833
- Iter 024 / 025, Loss: 0.390649

* Train accuracy / confusion: 83.25% / [[311, 29, 10], [43, 209, 21], [13, 18, 146]], 
* Val   accuracy / confusion: 52.88% / [[25, 12, 9], [9, 18, 8], [3, 8, 12]]

------------------------------ Epoch 263 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.266725
- Iter 024 / 025, Loss: 0.565816

* Train accuracy / confusion: 82.25% / [[312, 25, 17], [38, 209, 22], [22, 18, 137]], 
* Val   accuracy / confusion: 52.88% / [[31, 10, 5], [11, 16, 8], [8, 7, 8]]

------------------------------ Epoch 264 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.751757
- Iter 024 / 025, Loss: 0.324953

* Train accuracy / confusion: 84.25% / [[328, 21, 11], [35, 212, 20], [17, 22, 134]], 
* Val   accuracy / confusion: 60.58% / [[32, 9, 5], [6, 22, 7], [7, 7, 9]]

------------------------------ Epoch 265 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.452949
- Iter 024 / 025, Loss: 0.404571

* Train accuracy / confusion: 81.75% / [[317, 26, 14], [44, 203, 19], [23, 20, 134]], 
* Val   accuracy / confusion: 57.69% / [[29, 12, 5], [10, 18, 7], [5, 5, 13]]

------------------------------ Epoch 266 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.674791
- Iter 024 / 025, Loss: 0.636488

* Train accuracy / confusion: 82.50% / [[314, 27, 11], [44, 208, 20], [16, 22, 138]], 
* Val   accuracy / confusion: 56.73% / [[29, 12, 5], [7, 19, 9], [7, 5, 11]]

------------------------------ Epoch 267 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.357362
- Iter 024 / 025, Loss: 0.360296

* Train accuracy / confusion: 83.38% / [[317, 27, 12], [36, 214, 16], [21, 21, 136]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [10, 15, 10], [7, 6, 10]]

------------------------------ Epoch 268 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.437636
- Iter 024 / 025, Loss: 0.260144

* Train accuracy / confusion: 83.38% / [[319, 23, 12], [40, 212, 18], [17, 23, 136]], 
* Val   accuracy / confusion: 55.77% / [[30, 12, 4], [10, 14, 11], [5, 4, 14]]

------------------------------ Epoch 269 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.179246
- Iter 024 / 025, Loss: 0.341130

* Train accuracy / confusion: 80.62% / [[317, 31, 7], [47, 200, 21], [24, 25, 128]], 
* Val   accuracy / confusion: 50.00% / [[31, 10, 5], [10, 14, 11], [8, 8, 7]]

------------------------------ Epoch 270 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.597674
- Iter 024 / 025, Loss: 0.276428

* Train accuracy / confusion: 82.75% / [[318, 25, 10], [47, 200, 24], [15, 17, 144]], 
* Val   accuracy / confusion: 49.04% / [[30, 11, 5], [17, 14, 4], [9, 7, 7]]

------------------------------ Epoch 271 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.456720
- Iter 024 / 025, Loss: 0.386107

* Train accuracy / confusion: 82.88% / [[315, 28, 10], [36, 207, 26], [19, 18, 141]], 
* Val   accuracy / confusion: 48.08% / [[30, 11, 5], [17, 11, 7], [5, 9, 9]]

------------------------------ Epoch 272 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.436460
- Iter 024 / 025, Loss: 0.489542

* Train accuracy / confusion: 81.12% / [[309, 36, 12], [41, 213, 18], [19, 25, 127]], 
* Val   accuracy / confusion: 53.85% / [[31, 9, 6], [12, 16, 7], [7, 7, 9]]

------------------------------ Epoch 273 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.408671
- Iter 024 / 025, Loss: 0.245863

* Train accuracy / confusion: 83.25% / [[317, 30, 12], [28, 212, 28], [19, 17, 137]], 
* Val   accuracy / confusion: 55.77% / [[33, 6, 7], [11, 17, 7], [5, 10, 8]]

------------------------------ Epoch 274 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.345889
- Iter 024 / 025, Loss: 0.490563

* Train accuracy / confusion: 84.00% / [[324, 17, 13], [43, 206, 19], [15, 21, 142]], 
* Val   accuracy / confusion: 57.69% / [[33, 7, 6], [10, 19, 6], [7, 8, 8]]

------------------------------ Epoch 275 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.516354
- Iter 024 / 025, Loss: 0.281195

* Train accuracy / confusion: 82.75% / [[318, 24, 10], [43, 210, 17], [22, 22, 134]], 
* Val   accuracy / confusion: 50.96% / [[30, 11, 5], [10, 16, 9], [8, 8, 7]]

------------------------------ Epoch 276 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.438169
- Iter 024 / 025, Loss: 0.302556

* Train accuracy / confusion: 83.50% / [[321, 27, 7], [49, 203, 16], [15, 18, 144]], 
* Val   accuracy / confusion: 44.23% / [[29, 10, 7], [18, 12, 5], [10, 8, 5]]

------------------------------ Epoch 277 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.306037
- Iter 024 / 025, Loss: 0.415742

* Train accuracy / confusion: 84.75% / [[328, 20, 8], [42, 206, 24], [13, 15, 144]], 
* Val   accuracy / confusion: 50.96% / [[28, 13, 5], [10, 16, 9], [5, 9, 9]]

------------------------------ Epoch 278 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.391085
- Iter 024 / 025, Loss: 0.434761

* Train accuracy / confusion: 85.38% / [[327, 19, 11], [34, 218, 14], [18, 21, 138]], 
* Val   accuracy / confusion: 53.85% / [[30, 11, 5], [12, 16, 7], [8, 5, 10]]

------------------------------ Epoch 279 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.362465
- Iter 024 / 025, Loss: 0.302093

* Train accuracy / confusion: 83.12% / [[319, 24, 10], [43, 206, 19], [12, 27, 140]], 
* Val   accuracy / confusion: 61.54% / [[32, 11, 3], [5, 21, 9], [4, 8, 11]]

------------------------------ Epoch 280 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.360797
- Iter 024 / 025, Loss: 0.354747

* Train accuracy / confusion: 84.25% / [[327, 27, 6], [45, 206, 16], [19, 13, 141]], 
* Val   accuracy / confusion: 55.77% / [[31, 9, 6], [10, 16, 9], [3, 9, 11]]

------------------------------ Epoch 281 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.378895
- Iter 024 / 025, Loss: 0.379959

* Train accuracy / confusion: 83.25% / [[321, 23, 12], [34, 212, 21], [21, 23, 133]], 
* Val   accuracy / confusion: 52.88% / [[29, 11, 6], [11, 17, 7], [6, 8, 9]]

------------------------------ Epoch 282 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.422636
- Iter 024 / 025, Loss: 0.390998

* Train accuracy / confusion: 82.88% / [[313, 28, 15], [37, 206, 23], [15, 19, 144]], 
* Val   accuracy / confusion: 53.85% / [[28, 12, 6], [8, 17, 10], [6, 6, 11]]

------------------------------ Epoch 283 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.264395
- Iter 024 / 025, Loss: 0.236701

* Train accuracy / confusion: 85.12% / [[324, 22, 10], [25, 222, 20], [18, 24, 135]], 
* Val   accuracy / confusion: 52.88% / [[28, 12, 6], [8, 19, 8], [9, 6, 8]]

------------------------------ Epoch 284 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.429002
- Iter 024 / 025, Loss: 0.500906

* Train accuracy / confusion: 84.75% / [[327, 22, 5], [36, 212, 22], [18, 19, 139]], 
* Val   accuracy / confusion: 56.73% / [[31, 9, 6], [7, 19, 9], [7, 7, 9]]

------------------------------ Epoch 285 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.334239
- Iter 024 / 025, Loss: 0.357304

* Train accuracy / confusion: 82.50% / [[320, 23, 13], [40, 205, 23], [19, 22, 135]], 
* Val   accuracy / confusion: 54.81% / [[28, 14, 4], [10, 18, 7], [4, 8, 11]]

------------------------------ Epoch 286 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.382609
- Iter 024 / 025, Loss: 0.389856

* Train accuracy / confusion: 84.38% / [[313, 25, 15], [38, 215, 18], [10, 19, 147]], 
* Val   accuracy / confusion: 53.85% / [[32, 8, 6], [12, 15, 8], [7, 7, 9]]

------------------------------ Epoch 287 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.310256
- Iter 024 / 025, Loss: 0.574052

* Train accuracy / confusion: 84.12% / [[315, 26, 13], [41, 208, 17], [13, 17, 150]], 
* Val   accuracy / confusion: 51.92% / [[30, 13, 3], [12, 13, 10], [6, 6, 11]]

------------------------------ Epoch 288 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.590977
- Iter 024 / 025, Loss: 0.519214

* Train accuracy / confusion: 84.75% / [[324, 25, 10], [33, 214, 18], [16, 20, 140]], 
* Val   accuracy / confusion: 55.77% / [[29, 8, 9], [9, 19, 7], [7, 6, 10]]

------------------------------ Epoch 289 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.449984
- Iter 024 / 025, Loss: 0.146340

* Train accuracy / confusion: 85.50% / [[324, 16, 14], [35, 216, 17], [15, 19, 144]], 
* Val   accuracy / confusion: 56.73% / [[32, 8, 6], [10, 17, 8], [5, 8, 10]]

------------------------------ Epoch 290 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.224274
- Iter 024 / 025, Loss: 0.332504

* Train accuracy / confusion: 85.50% / [[331, 21, 5], [32, 210, 28], [8, 22, 143]], 
* Val   accuracy / confusion: 53.85% / [[30, 11, 5], [10, 16, 9], [7, 6, 10]]

------------------------------ Epoch 291 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.473360
- Iter 024 / 025, Loss: 0.552613

* Train accuracy / confusion: 84.50% / [[328, 22, 8], [32, 212, 21], [17, 24, 136]], 
* Val   accuracy / confusion: 51.92% / [[33, 9, 4], [14, 15, 6], [6, 11, 6]]

------------------------------ Epoch 292 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.451729
- Iter 024 / 025, Loss: 0.300738

* Train accuracy / confusion: 84.88% / [[325, 25, 7], [35, 219, 17], [17, 20, 135]], 
* Val   accuracy / confusion: 54.81% / [[32, 7, 7], [10, 16, 9], [5, 9, 9]]

------------------------------ Epoch 293 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.479532
- Iter 024 / 025, Loss: 0.551792

* Train accuracy / confusion: 82.38% / [[324, 23, 13], [42, 201, 24], [12, 27, 134]], 
* Val   accuracy / confusion: 48.08% / [[25, 10, 11], [10, 14, 11], [6, 6, 11]]

------------------------------ Epoch 294 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.486770
- Iter 024 / 025, Loss: 0.320477

* Train accuracy / confusion: 83.38% / [[326, 26, 6], [42, 202, 23], [13, 23, 139]], 
* Val   accuracy / confusion: 48.08% / [[26, 14, 6], [12, 15, 8], [8, 6, 9]]

------------------------------ Epoch 295 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.285332
- Iter 024 / 025, Loss: 0.404697

* Train accuracy / confusion: 83.75% / [[329, 22, 7], [41, 205, 22], [14, 24, 136]], 
* Val   accuracy / confusion: 53.85% / [[30, 11, 5], [13, 15, 7], [6, 6, 11]]

------------------------------ Epoch 296 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.240116
- Iter 024 / 025, Loss: 0.483288

* Train accuracy / confusion: 86.25% / [[327, 19, 9], [38, 217, 14], [12, 18, 146]], 
* Val   accuracy / confusion: 49.04% / [[28, 13, 5], [12, 16, 7], [7, 9, 7]]

------------------------------ Epoch 297 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.400836
- Iter 024 / 025, Loss: 0.271488

* Train accuracy / confusion: 82.62% / [[307, 31, 16], [37, 211, 22], [12, 21, 143]], 
* Val   accuracy / confusion: 50.00% / [[29, 11, 6], [11, 14, 10], [9, 5, 9]]

------------------------------ Epoch 298 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.396511
- Iter 024 / 025, Loss: 0.316730

* Train accuracy / confusion: 84.38% / [[323, 15, 15], [39, 209, 21], [16, 19, 143]], 
* Val   accuracy / confusion: 47.12% / [[24, 15, 7], [10, 16, 9], [9, 5, 9]]

------------------------------ Epoch 299 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.518383
- Iter 024 / 025, Loss: 0.268235

* Train accuracy / confusion: 86.12% / [[320, 25, 6], [37, 220, 17], [11, 15, 149]], 
* Val   accuracy / confusion: 52.88% / [[31, 9, 6], [10, 17, 8], [7, 9, 7]]

------------------------------ Epoch 300 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.455522
- Iter 024 / 025, Loss: 0.576788

* Train accuracy / confusion: 85.75% / [[324, 16, 12], [32, 221, 18], [12, 24, 141]], 
* Val   accuracy / confusion: 57.69% / [[31, 9, 6], [8, 17, 10], [6, 5, 12]]

------------------------------ Epoch 301 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.348799
- Iter 024 / 025, Loss: 0.318613

* Train accuracy / confusion: 84.50% / [[320, 30, 6], [41, 213, 16], [15, 16, 143]], 
* Val   accuracy / confusion: 48.08% / [[30, 9, 7], [11, 12, 12], [9, 6, 8]]

------------------------------ Epoch 302 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.172088
- Iter 024 / 025, Loss: 0.268350

* Train accuracy / confusion: 84.12% / [[326, 24, 5], [37, 209, 26], [15, 20, 138]], 
* Val   accuracy / confusion: 54.81% / [[31, 12, 3], [8, 18, 9], [10, 5, 8]]

------------------------------ Epoch 303 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.568227
- Iter 024 / 025, Loss: 0.139650

* Train accuracy / confusion: 83.88% / [[321, 24, 9], [32, 209, 27], [11, 26, 141]], 
* Val   accuracy / confusion: 50.96% / [[33, 5, 8], [9, 14, 12], [8, 9, 6]]

------------------------------ Epoch 304 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.389002
- Iter 024 / 025, Loss: 0.282218

* Train accuracy / confusion: 85.25% / [[318, 31, 8], [33, 220, 17], [13, 16, 144]], 
* Val   accuracy / confusion: 54.81% / [[31, 9, 6], [10, 16, 9], [7, 6, 10]]

------------------------------ Epoch 305 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.250626
- Iter 024 / 025, Loss: 0.460066

* Train accuracy / confusion: 84.62% / [[321, 23, 12], [36, 213, 22], [11, 19, 143]], 
* Val   accuracy / confusion: 53.85% / [[30, 9, 7], [12, 13, 10], [2, 8, 13]]

------------------------------ Epoch 306 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.369028
- Iter 024 / 025, Loss: 0.247305

* Train accuracy / confusion: 84.50% / [[326, 23, 11], [37, 213, 18], [15, 20, 137]], 
* Val   accuracy / confusion: 53.85% / [[32, 9, 5], [14, 14, 7], [6, 7, 10]]

------------------------------ Epoch 307 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.373247
- Iter 024 / 025, Loss: 0.313901

* Train accuracy / confusion: 84.50% / [[326, 16, 10], [37, 212, 20], [20, 21, 138]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [9, 17, 9], [6, 9, 8]]

------------------------------ Epoch 308 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.372058
- Iter 024 / 025, Loss: 0.216184

* Train accuracy / confusion: 86.75% / [[323, 22, 9], [33, 221, 13], [12, 17, 150]], 
* Val   accuracy / confusion: 50.00% / [[29, 10, 7], [13, 13, 9], [8, 5, 10]]

------------------------------ Epoch 309 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.232620
- Iter 024 / 025, Loss: 0.181812

* Train accuracy / confusion: 86.00% / [[324, 22, 12], [34, 213, 19], [9, 16, 151]], 
* Val   accuracy / confusion: 53.85% / [[32, 9, 5], [10, 14, 11], [3, 10, 10]]

------------------------------ Epoch 310 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.403198
- Iter 024 / 025, Loss: 0.318204

* Train accuracy / confusion: 85.38% / [[327, 19, 7], [29, 218, 21], [17, 24, 138]], 
* Val   accuracy / confusion: 51.92% / [[29, 9, 8], [9, 18, 8], [8, 8, 7]]

------------------------------ Epoch 311 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.374241
- Iter 024 / 025, Loss: 0.361888

* Train accuracy / confusion: 82.88% / [[312, 36, 16], [37, 208, 19], [13, 16, 143]], 
* Val   accuracy / confusion: 55.77% / [[33, 8, 5], [15, 14, 6], [6, 6, 11]]

------------------------------ Epoch 312 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.340439
- Iter 024 / 025, Loss: 0.312784

* Train accuracy / confusion: 84.00% / [[326, 24, 5], [48, 199, 23], [10, 18, 147]], 
* Val   accuracy / confusion: 50.96% / [[28, 12, 6], [10, 15, 10], [7, 6, 10]]

------------------------------ Epoch 313 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.149113
- Iter 024 / 025, Loss: 0.339230

* Train accuracy / confusion: 86.62% / [[321, 23, 9], [26, 229, 15], [18, 16, 143]], 
* Val   accuracy / confusion: 51.92% / [[30, 7, 9], [12, 18, 5], [6, 11, 6]]

------------------------------ Epoch 314 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.425186
- Iter 024 / 025, Loss: 0.268110

* Train accuracy / confusion: 85.25% / [[316, 23, 16], [31, 223, 15], [18, 15, 143]], 
* Val   accuracy / confusion: 54.81% / [[26, 12, 8], [8, 19, 8], [5, 6, 12]]

------------------------------ Epoch 315 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.512596
- Iter 024 / 025, Loss: 0.214182

* Train accuracy / confusion: 83.62% / [[316, 22, 19], [37, 216, 18], [15, 20, 137]], 
* Val   accuracy / confusion: 52.88% / [[32, 6, 8], [13, 13, 9], [7, 6, 10]]

------------------------------ Epoch 316 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.229652
- Iter 024 / 025, Loss: 0.524142

* Train accuracy / confusion: 84.88% / [[315, 28, 11], [34, 218, 18], [13, 17, 146]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [11, 13, 11], [6, 5, 12]]

------------------------------ Epoch 317 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.504968
- Iter 024 / 025, Loss: 0.414762

* Train accuracy / confusion: 87.25% / [[328, 28, 5], [31, 216, 18], [6, 14, 154]], 
* Val   accuracy / confusion: 49.04% / [[29, 9, 8], [13, 15, 7], [8, 8, 7]]

------------------------------ Epoch 318 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.319118
- Iter 024 / 025, Loss: 0.426208

* Train accuracy / confusion: 85.12% / [[327, 24, 6], [31, 211, 25], [14, 19, 143]], 
* Val   accuracy / confusion: 51.92% / [[29, 9, 8], [13, 15, 7], [7, 6, 10]]

------------------------------ Epoch 319 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.592540
- Iter 024 / 025, Loss: 0.388810

* Train accuracy / confusion: 86.38% / [[324, 22, 9], [38, 218, 14], [11, 15, 149]], 
* Val   accuracy / confusion: 47.12% / [[27, 12, 7], [9, 15, 11], [8, 8, 7]]

------------------------------ Epoch 320 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.435296
- Iter 024 / 025, Loss: 0.239230

* Train accuracy / confusion: 82.50% / [[322, 26, 13], [36, 199, 29], [16, 20, 139]], 
* Val   accuracy / confusion: 51.92% / [[28, 12, 6], [12, 14, 9], [4, 7, 12]]

------------------------------ Epoch 321 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.443526
- Iter 024 / 025, Loss: 0.414085

* Train accuracy / confusion: 86.50% / [[319, 34, 5], [26, 225, 16], [14, 13, 148]], 
* Val   accuracy / confusion: 56.73% / [[32, 9, 5], [10, 20, 5], [7, 9, 7]]

------------------------------ Epoch 322 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.371892
- Iter 024 / 025, Loss: 0.318527

* Train accuracy / confusion: 84.12% / [[313, 29, 13], [31, 219, 17], [15, 22, 141]], 
* Val   accuracy / confusion: 56.73% / [[33, 9, 4], [8, 16, 11], [6, 7, 10]]

------------------------------ Epoch 323 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.463039
- Iter 024 / 025, Loss: 0.300304

* Train accuracy / confusion: 85.00% / [[322, 21, 12], [41, 216, 12], [13, 21, 142]], 
* Val   accuracy / confusion: 57.69% / [[32, 8, 6], [9, 19, 7], [7, 7, 9]]

------------------------------ Epoch 324 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.297113
- Iter 024 / 025, Loss: 0.488215

* Train accuracy / confusion: 83.38% / [[318, 27, 8], [36, 211, 22], [16, 24, 138]], 
* Val   accuracy / confusion: 57.69% / [[33, 9, 4], [14, 17, 4], [5, 8, 10]]

------------------------------ Epoch 325 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.458622
- Iter 024 / 025, Loss: 0.390430

* Train accuracy / confusion: 85.00% / [[317, 27, 9], [46, 213, 14], [12, 12, 150]], 
* Val   accuracy / confusion: 60.58% / [[31, 8, 7], [7, 16, 12], [6, 1, 16]]

------------------------------ Epoch 326 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.462906
- Iter 024 / 025, Loss: 0.334867

* Train accuracy / confusion: 86.75% / [[332, 15, 7], [35, 215, 20], [18, 11, 147]], 
* Val   accuracy / confusion: 58.65% / [[28, 13, 5], [11, 20, 4], [6, 4, 13]]

------------------------------ Epoch 327 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.250893
- Iter 024 / 025, Loss: 0.399422

* Train accuracy / confusion: 86.00% / [[329, 15, 10], [33, 214, 23], [19, 12, 145]], 
* Val   accuracy / confusion: 60.58% / [[30, 9, 7], [8, 20, 7], [2, 8, 13]]

------------------------------ Epoch 328 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.270975
- Iter 024 / 025, Loss: 0.294745

* Train accuracy / confusion: 84.62% / [[322, 22, 15], [36, 213, 15], [15, 20, 142]], 
* Val   accuracy / confusion: 50.96% / [[27, 12, 7], [11, 16, 8], [6, 7, 10]]

------------------------------ Epoch 329 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.287784
- Iter 024 / 025, Loss: 0.265480

* Train accuracy / confusion: 85.12% / [[321, 23, 10], [36, 215, 20], [15, 15, 145]], 
* Val   accuracy / confusion: 53.85% / [[27, 11, 8], [8, 21, 6], [3, 12, 8]]

------------------------------ Epoch 330 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.214742
- Iter 024 / 025, Loss: 0.544588

* Train accuracy / confusion: 84.12% / [[309, 33, 11], [36, 218, 18], [12, 17, 146]], 
* Val   accuracy / confusion: 48.08% / [[28, 13, 5], [10, 13, 12], [6, 8, 9]]

------------------------------ Epoch 331 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.332782
- Iter 024 / 025, Loss: 0.309291

* Train accuracy / confusion: 86.38% / [[326, 16, 8], [34, 215, 21], [16, 14, 150]], 
* Val   accuracy / confusion: 56.73% / [[30, 13, 3], [9, 20, 6], [8, 6, 9]]

------------------------------ Epoch 332 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.301397
- Iter 024 / 025, Loss: 0.260410

* Train accuracy / confusion: 84.88% / [[322, 26, 10], [33, 213, 21], [13, 18, 144]], 
* Val   accuracy / confusion: 58.65% / [[28, 12, 6], [6, 22, 7], [3, 9, 11]]

------------------------------ Epoch 333 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.291619
- Iter 024 / 025, Loss: 0.279596

* Train accuracy / confusion: 85.00% / [[317, 27, 11], [37, 219, 17], [16, 12, 144]], 
* Val   accuracy / confusion: 55.77% / [[33, 7, 6], [12, 14, 9], [9, 3, 11]]

------------------------------ Epoch 334 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.451820
- Iter 024 / 025, Loss: 0.249924

* Train accuracy / confusion: 85.62% / [[323, 23, 11], [38, 214, 19], [14, 10, 148]], 
* Val   accuracy / confusion: 54.81% / [[32, 7, 7], [11, 19, 5], [10, 7, 6]]

------------------------------ Epoch 335 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.353084
- Iter 024 / 025, Loss: 0.491059

* Train accuracy / confusion: 85.38% / [[322, 27, 9], [27, 220, 18], [15, 21, 141]], 
* Val   accuracy / confusion: 47.12% / [[27, 15, 4], [12, 12, 11], [8, 5, 10]]

------------------------------ Epoch 336 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.110853
- Iter 024 / 025, Loss: 0.432319

* Train accuracy / confusion: 86.75% / [[321, 20, 14], [30, 229, 12], [11, 19, 144]], 
* Val   accuracy / confusion: 61.54% / [[30, 11, 5], [6, 21, 8], [5, 5, 13]]

------------------------------ Epoch 337 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.355410
- Iter 024 / 025, Loss: 0.676247

* Train accuracy / confusion: 86.75% / [[328, 20, 9], [33, 218, 15], [11, 18, 148]], 
* Val   accuracy / confusion: 51.92% / [[29, 12, 5], [13, 16, 6], [5, 9, 9]]

------------------------------ Epoch 338 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.287945
- Iter 024 / 025, Loss: 0.203161

* Train accuracy / confusion: 85.00% / [[320, 28, 8], [32, 215, 20], [16, 16, 145]], 
* Val   accuracy / confusion: 48.08% / [[27, 12, 7], [13, 15, 7], [5, 10, 8]]

------------------------------ Epoch 339 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.207096
- Iter 024 / 025, Loss: 0.249538

* Train accuracy / confusion: 87.25% / [[328, 25, 2], [38, 220, 15], [10, 12, 150]], 
* Val   accuracy / confusion: 50.96% / [[29, 13, 4], [14, 13, 8], [5, 7, 11]]

------------------------------ Epoch 340 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.297772
- Iter 024 / 025, Loss: 0.384861

* Train accuracy / confusion: 86.38% / [[332, 18, 6], [36, 216, 17], [12, 20, 143]], 
* Val   accuracy / confusion: 51.92% / [[29, 10, 7], [12, 15, 8], [4, 9, 10]]

------------------------------ Epoch 341 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.312316
- Iter 024 / 025, Loss: 0.524503

* Train accuracy / confusion: 84.25% / [[331, 19, 9], [31, 212, 26], [14, 27, 131]], 
* Val   accuracy / confusion: 50.96% / [[27, 12, 7], [11, 16, 8], [7, 6, 10]]

------------------------------ Epoch 342 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.425667
- Iter 024 / 025, Loss: 0.344912

* Train accuracy / confusion: 86.75% / [[322, 22, 11], [28, 226, 17], [15, 13, 146]], 
* Val   accuracy / confusion: 56.73% / [[31, 9, 6], [9, 18, 8], [5, 8, 10]]

------------------------------ Epoch 343 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.263904
- Iter 024 / 025, Loss: 0.338912

* Train accuracy / confusion: 86.75% / [[326, 23, 10], [29, 219, 18], [12, 14, 149]], 
* Val   accuracy / confusion: 49.04% / [[25, 12, 9], [12, 13, 10], [5, 5, 13]]

------------------------------ Epoch 344 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.314771
- Iter 024 / 025, Loss: 0.392600

* Train accuracy / confusion: 86.25% / [[332, 19, 7], [36, 216, 15], [19, 14, 142]], 
* Val   accuracy / confusion: 49.04% / [[27, 13, 6], [10, 16, 9], [7, 8, 8]]

------------------------------ Epoch 345 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.656425
- Iter 024 / 025, Loss: 0.604947

* Train accuracy / confusion: 85.25% / [[328, 23, 10], [34, 210, 20], [13, 18, 144]], 
* Val   accuracy / confusion: 49.04% / [[28, 10, 8], [9, 16, 10], [8, 8, 7]]

------------------------------ Epoch 346 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.447677
- Iter 024 / 025, Loss: 0.309470

* Train accuracy / confusion: 85.50% / [[318, 27, 11], [35, 219, 13], [15, 15, 147]], 
* Val   accuracy / confusion: 44.23% / [[22, 17, 7], [15, 16, 4], [4, 11, 8]]

------------------------------ Epoch 347 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.346009
- Iter 024 / 025, Loss: 0.628388

* Train accuracy / confusion: 86.75% / [[325, 19, 6], [32, 218, 20], [14, 15, 151]], 
* Val   accuracy / confusion: 50.00% / [[29, 9, 8], [8, 13, 14], [4, 9, 10]]

------------------------------ Epoch 348 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.303520
- Iter 024 / 025, Loss: 0.408690

* Train accuracy / confusion: 87.88% / [[328, 20, 7], [30, 223, 15], [14, 11, 152]], 
* Val   accuracy / confusion: 57.69% / [[31, 8, 7], [9, 18, 8], [4, 8, 11]]

------------------------------ Epoch 349 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.409956
- Iter 024 / 025, Loss: 0.188339

* Train accuracy / confusion: 88.00% / [[331, 16, 7], [19, 224, 23], [9, 22, 149]], 
* Val   accuracy / confusion: 49.04% / [[25, 13, 8], [10, 17, 8], [9, 5, 9]]

------------------------------ Epoch 350 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.674138
- Iter 024 / 025, Loss: 0.196185

* Train accuracy / confusion: 86.75% / [[322, 23, 11], [28, 223, 16], [12, 16, 149]], 
* Val   accuracy / confusion: 52.88% / [[29, 10, 7], [12, 14, 9], [5, 6, 12]]

------------------------------ Epoch 351 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.540698
- Iter 024 / 025, Loss: 0.281496

* Train accuracy / confusion: 83.38% / [[325, 21, 13], [42, 206, 22], [16, 19, 136]], 
* Val   accuracy / confusion: 48.08% / [[28, 13, 5], [10, 15, 10], [5, 11, 7]]

------------------------------ Epoch 352 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.282991
- Iter 024 / 025, Loss: 0.154589

* Train accuracy / confusion: 88.12% / [[321, 24, 7], [27, 226, 17], [8, 12, 158]], 
* Val   accuracy / confusion: 51.92% / [[27, 10, 9], [13, 15, 7], [5, 6, 12]]

------------------------------ Epoch 353 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.314504
- Iter 024 / 025, Loss: 0.202471

* Train accuracy / confusion: 87.75% / [[331, 14, 10], [35, 219, 14], [9, 16, 152]], 
* Val   accuracy / confusion: 50.00% / [[30, 8, 8], [10, 16, 9], [7, 10, 6]]

------------------------------ Epoch 354 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.386349
- Iter 024 / 025, Loss: 0.598981

* Train accuracy / confusion: 84.88% / [[319, 25, 14], [35, 214, 17], [12, 18, 146]], 
* Val   accuracy / confusion: 46.15% / [[24, 16, 6], [14, 17, 4], [4, 12, 7]]

------------------------------ Epoch 355 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.440942
- Iter 024 / 025, Loss: 0.369584

* Train accuracy / confusion: 85.75% / [[317, 27, 8], [34, 221, 13], [13, 19, 148]], 
* Val   accuracy / confusion: 50.00% / [[29, 12, 5], [12, 14, 9], [5, 9, 9]]

------------------------------ Epoch 356 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.172703
- Iter 024 / 025, Loss: 0.380410

* Train accuracy / confusion: 88.38% / [[336, 13, 7], [23, 227, 22], [6, 22, 144]], 
* Val   accuracy / confusion: 52.88% / [[27, 11, 8], [8, 17, 10], [6, 6, 11]]

------------------------------ Epoch 357 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.463770
- Iter 024 / 025, Loss: 0.424409

* Train accuracy / confusion: 86.75% / [[331, 15, 7], [26, 220, 23], [17, 18, 143]], 
* Val   accuracy / confusion: 52.88% / [[26, 15, 5], [9, 19, 7], [4, 9, 10]]

------------------------------ Epoch 358 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.684925
- Iter 024 / 025, Loss: 0.496308

* Train accuracy / confusion: 84.00% / [[323, 27, 9], [35, 211, 19], [16, 22, 138]], 
* Val   accuracy / confusion: 45.19% / [[26, 12, 8], [14, 12, 9], [7, 7, 9]]

------------------------------ Epoch 359 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.340658
- Iter 024 / 025, Loss: 0.204272

* Train accuracy / confusion: 88.12% / [[334, 16, 10], [25, 229, 15], [10, 19, 142]], 
* Val   accuracy / confusion: 47.12% / [[27, 12, 7], [12, 12, 11], [9, 4, 10]]

------------------------------ Epoch 360 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.284478
- Iter 024 / 025, Loss: 0.232482

* Train accuracy / confusion: 85.50% / [[325, 20, 6], [39, 211, 24], [11, 16, 148]], 
* Val   accuracy / confusion: 54.81% / [[27, 8, 11], [7, 18, 10], [7, 4, 12]]

------------------------------ Epoch 361 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.417886
- Iter 024 / 025, Loss: 0.322669

* Train accuracy / confusion: 84.62% / [[315, 29, 13], [35, 213, 21], [12, 13, 149]], 
* Val   accuracy / confusion: 47.12% / [[30, 10, 6], [14, 13, 8], [6, 11, 6]]

------------------------------ Epoch 362 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.236185
- Iter 024 / 025, Loss: 0.439528

* Train accuracy / confusion: 86.25% / [[324, 22, 8], [31, 226, 14], [12, 23, 140]], 
* Val   accuracy / confusion: 51.92% / [[30, 8, 8], [12, 14, 9], [6, 7, 10]]

------------------------------ Epoch 363 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.439820
- Iter 024 / 025, Loss: 0.356745

* Train accuracy / confusion: 86.50% / [[332, 17, 8], [28, 216, 23], [9, 23, 144]], 
* Val   accuracy / confusion: 48.08% / [[25, 16, 5], [11, 16, 8], [7, 7, 9]]

------------------------------ Epoch 364 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.294753
- Iter 024 / 025, Loss: 0.435906

* Train accuracy / confusion: 85.38% / [[321, 18, 12], [36, 211, 25], [9, 17, 151]], 
* Val   accuracy / confusion: 51.92% / [[27, 8, 11], [7, 18, 10], [7, 7, 9]]

------------------------------ Epoch 365 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.312047
- Iter 024 / 025, Loss: 0.263082

* Train accuracy / confusion: 86.25% / [[325, 18, 15], [30, 221, 16], [15, 16, 144]], 
* Val   accuracy / confusion: 51.92% / [[26, 11, 9], [10, 17, 8], [4, 8, 11]]

------------------------------ Epoch 366 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.263348
- Iter 024 / 025, Loss: 0.351112

* Train accuracy / confusion: 87.75% / [[327, 23, 9], [27, 229, 11], [12, 16, 146]], 
* Val   accuracy / confusion: 53.85% / [[29, 10, 7], [12, 17, 6], [7, 6, 10]]

------------------------------ Epoch 367 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.383355
- Iter 024 / 025, Loss: 0.283479

* Train accuracy / confusion: 87.50% / [[330, 16, 7], [37, 222, 11], [12, 17, 148]], 
* Val   accuracy / confusion: 54.81% / [[30, 11, 5], [12, 18, 5], [5, 9, 9]]

------------------------------ Epoch 368 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.335731
- Iter 024 / 025, Loss: 0.351034

* Train accuracy / confusion: 87.00% / [[321, 18, 16], [27, 228, 15], [11, 17, 147]], 
* Val   accuracy / confusion: 56.73% / [[28, 10, 8], [10, 18, 7], [2, 8, 13]]

------------------------------ Epoch 369 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.402914
- Iter 024 / 025, Loss: 0.355983

* Train accuracy / confusion: 85.75% / [[322, 25, 11], [30, 222, 16], [13, 19, 142]], 
* Val   accuracy / confusion: 53.85% / [[26, 13, 7], [8, 20, 7], [7, 6, 10]]

------------------------------ Epoch 370 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.368666
- Iter 024 / 025, Loss: 0.303897

* Train accuracy / confusion: 86.50% / [[330, 19, 11], [33, 220, 15], [8, 22, 142]], 
* Val   accuracy / confusion: 52.88% / [[29, 10, 7], [12, 16, 7], [6, 7, 10]]

------------------------------ Epoch 371 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.289681
- Iter 024 / 025, Loss: 0.126122

* Train accuracy / confusion: 88.38% / [[333, 16, 7], [27, 226, 15], [12, 16, 148]], 
* Val   accuracy / confusion: 50.96% / [[27, 12, 7], [10, 19, 6], [4, 12, 7]]

------------------------------ Epoch 372 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.369513
- Iter 024 / 025, Loss: 0.403806

* Train accuracy / confusion: 87.25% / [[331, 16, 6], [41, 212, 15], [10, 14, 155]], 
* Val   accuracy / confusion: 55.77% / [[32, 7, 7], [11, 16, 8], [6, 7, 10]]

------------------------------ Epoch 373 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.336827
- Iter 024 / 025, Loss: 0.325705

* Train accuracy / confusion: 87.50% / [[333, 15, 10], [34, 219, 19], [8, 14, 148]], 
* Val   accuracy / confusion: 57.69% / [[33, 11, 2], [10, 19, 6], [6, 9, 8]]

------------------------------ Epoch 374 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.258595
- Iter 024 / 025, Loss: 0.421164

* Train accuracy / confusion: 86.25% / [[322, 26, 8], [29, 224, 22], [12, 13, 144]], 
* Val   accuracy / confusion: 54.81% / [[28, 11, 7], [7, 21, 7], [5, 10, 8]]

------------------------------ Epoch 375 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.186165
- Iter 024 / 025, Loss: 0.448084

* Train accuracy / confusion: 88.50% / [[340, 14, 5], [32, 224, 11], [14, 16, 144]], 
* Val   accuracy / confusion: 57.69% / [[31, 8, 7], [11, 19, 5], [6, 7, 10]]

------------------------------ Epoch 376 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.441157
- Iter 024 / 025, Loss: 0.172701

* Train accuracy / confusion: 86.50% / [[325, 22, 8], [27, 226, 18], [14, 19, 141]], 
* Val   accuracy / confusion: 54.81% / [[31, 10, 5], [10, 16, 9], [4, 9, 10]]

------------------------------ Epoch 377 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.240600
- Iter 024 / 025, Loss: 0.495166

* Train accuracy / confusion: 86.75% / [[326, 20, 10], [29, 224, 17], [12, 18, 144]], 
* Val   accuracy / confusion: 61.54% / [[31, 9, 6], [6, 21, 8], [4, 7, 12]]

------------------------------ Epoch 378 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.494785
- Iter 024 / 025, Loss: 0.487454

* Train accuracy / confusion: 85.75% / [[324, 24, 8], [30, 216, 20], [10, 22, 146]], 
* Val   accuracy / confusion: 54.81% / [[32, 6, 8], [11, 16, 8], [8, 6, 9]]

------------------------------ Epoch 379 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.292074
- Iter 024 / 025, Loss: 0.242720

* Train accuracy / confusion: 86.25% / [[326, 22, 9], [31, 219, 20], [13, 15, 145]], 
* Val   accuracy / confusion: 50.96% / [[24, 15, 7], [8, 18, 9], [6, 6, 11]]

------------------------------ Epoch 380 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.425788
- Iter 024 / 025, Loss: 0.250202

* Train accuracy / confusion: 86.00% / [[329, 19, 9], [33, 221, 15], [15, 21, 138]], 
* Val   accuracy / confusion: 52.88% / [[28, 14, 4], [10, 15, 10], [5, 6, 12]]

------------------------------ Epoch 381 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.223651
- Iter 024 / 025, Loss: 0.366839

* Train accuracy / confusion: 86.88% / [[328, 22, 9], [31, 224, 13], [11, 19, 143]], 
* Val   accuracy / confusion: 47.12% / [[27, 11, 8], [11, 13, 11], [8, 6, 9]]

------------------------------ Epoch 382 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.403815
- Iter 024 / 025, Loss: 0.281673

* Train accuracy / confusion: 86.75% / [[327, 22, 10], [29, 219, 17], [13, 15, 148]], 
* Val   accuracy / confusion: 50.00% / [[27, 13, 6], [15, 13, 7], [7, 4, 12]]

------------------------------ Epoch 383 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.216305
- Iter 024 / 025, Loss: 0.287273

* Train accuracy / confusion: 87.62% / [[330, 21, 9], [31, 223, 13], [9, 16, 148]], 
* Val   accuracy / confusion: 50.96% / [[28, 12, 6], [12, 18, 5], [6, 10, 7]]

------------------------------ Epoch 384 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.377054
- Iter 024 / 025, Loss: 0.478411

* Train accuracy / confusion: 84.88% / [[326, 23, 8], [31, 216, 21], [19, 19, 137]], 
* Val   accuracy / confusion: 50.96% / [[29, 9, 8], [11, 15, 9], [8, 6, 9]]

------------------------------ Epoch 385 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.441508
- Iter 024 / 025, Loss: 0.302073

* Train accuracy / confusion: 88.25% / [[329, 22, 6], [25, 229, 12], [14, 15, 148]], 
* Val   accuracy / confusion: 53.85% / [[27, 10, 9], [8, 17, 10], [5, 6, 12]]

------------------------------ Epoch 386 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.426348
- Iter 024 / 025, Loss: 0.288370

* Train accuracy / confusion: 87.88% / [[327, 20, 10], [32, 225, 12], [10, 13, 151]], 
* Val   accuracy / confusion: 50.96% / [[33, 9, 4], [11, 16, 8], [5, 14, 4]]

------------------------------ Epoch 387 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.271532
- Iter 024 / 025, Loss: 0.551365

* Train accuracy / confusion: 87.75% / [[331, 20, 7], [30, 221, 13], [10, 18, 150]], 
* Val   accuracy / confusion: 50.00% / [[28, 10, 8], [13, 13, 9], [7, 5, 11]]

------------------------------ Epoch 388 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.373250
- Iter 024 / 025, Loss: 0.274531

* Train accuracy / confusion: 87.25% / [[330, 17, 9], [24, 221, 22], [13, 17, 147]], 
* Val   accuracy / confusion: 52.88% / [[26, 16, 4], [10, 17, 8], [3, 8, 12]]

------------------------------ Epoch 389 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.131414
- Iter 024 / 025, Loss: 0.219078

* Train accuracy / confusion: 87.50% / [[330, 15, 9], [26, 225, 20], [15, 15, 145]], 
* Val   accuracy / confusion: 50.00% / [[29, 10, 7], [11, 15, 9], [7, 8, 8]]

------------------------------ Epoch 390 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.324545
- Iter 024 / 025, Loss: 0.278554

* Train accuracy / confusion: 88.62% / [[328, 20, 10], [21, 232, 15], [10, 15, 149]], 
* Val   accuracy / confusion: 46.15% / [[27, 11, 8], [12, 12, 11], [6, 8, 9]]

------------------------------ Epoch 391 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.192340
- Iter 024 / 025, Loss: 0.281925

* Train accuracy / confusion: 87.38% / [[323, 20, 11], [26, 228, 17], [17, 10, 148]], 
* Val   accuracy / confusion: 48.08% / [[26, 13, 7], [11, 15, 9], [6, 8, 9]]

------------------------------ Epoch 392 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.200439
- Iter 024 / 025, Loss: 0.194864

* Train accuracy / confusion: 86.88% / [[328, 22, 8], [28, 218, 20], [10, 17, 149]], 
* Val   accuracy / confusion: 58.65% / [[31, 8, 7], [9, 20, 6], [6, 7, 10]]

------------------------------ Epoch 393 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.137019
- Iter 024 / 025, Loss: 0.432740

* Train accuracy / confusion: 86.62% / [[324, 25, 7], [34, 222, 10], [12, 19, 147]], 
* Val   accuracy / confusion: 57.69% / [[32, 10, 4], [9, 15, 11], [6, 4, 13]]

------------------------------ Epoch 394 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.201628
- Iter 024 / 025, Loss: 0.215338

* Train accuracy / confusion: 88.50% / [[332, 21, 7], [26, 225, 12], [12, 14, 151]], 
* Val   accuracy / confusion: 48.08% / [[26, 13, 7], [14, 14, 7], [7, 6, 10]]

------------------------------ Epoch 395 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.280935
- Iter 024 / 025, Loss: 0.219174

* Train accuracy / confusion: 86.88% / [[328, 16, 12], [27, 223, 20], [15, 15, 144]], 
* Val   accuracy / confusion: 53.85% / [[27, 12, 7], [11, 19, 5], [6, 7, 10]]

------------------------------ Epoch 396 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.264108
- Iter 024 / 025, Loss: 0.368723

* Train accuracy / confusion: 86.25% / [[319, 19, 12], [34, 218, 20], [11, 14, 153]], 
* Val   accuracy / confusion: 54.81% / [[32, 7, 7], [9, 16, 10], [7, 7, 9]]

------------------------------ Epoch 397 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.385113
- Iter 024 / 025, Loss: 0.279340

* Train accuracy / confusion: 86.75% / [[327, 24, 8], [26, 221, 19], [13, 16, 146]], 
* Val   accuracy / confusion: 52.88% / [[31, 9, 6], [13, 17, 5], [5, 11, 7]]

------------------------------ Epoch 398 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.281545
- Iter 024 / 025, Loss: 0.137028

* Train accuracy / confusion: 85.50% / [[313, 32, 10], [32, 226, 14], [11, 17, 145]], 
* Val   accuracy / confusion: 51.92% / [[29, 14, 3], [15, 15, 5], [4, 9, 10]]

------------------------------ Epoch 399 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.327676
- Iter 024 / 025, Loss: 0.302482

* Train accuracy / confusion: 88.25% / [[330, 20, 4], [31, 224, 13], [8, 18, 152]], 
* Val   accuracy / confusion: 53.85% / [[29, 12, 5], [11, 17, 7], [3, 10, 10]]

------------------------------ Epoch 400 / 500, Learning rate: 3.81e-04 ------------------------------
- Iter 012 / 025, Loss: 0.221045
- Iter 024 / 025, Loss: 0.179872

* Train accuracy / confusion: 88.25% / [[329, 20, 6], [27, 225, 15], [14, 12, 152]], 
* Val   accuracy / confusion: 50.00% / [[26, 13, 7], [10, 15, 10], [4, 8, 11]]

------------------------------ Epoch 401 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.307779
- Iter 024 / 025, Loss: 0.412217

* Train accuracy / confusion: 88.12% / [[329, 17, 6], [39, 223, 8], [9, 16, 153]], 
* Val   accuracy / confusion: 52.88% / [[32, 6, 8], [17, 11, 7], [5, 6, 12]]

------------------------------ Epoch 402 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.234723
- Iter 024 / 025, Loss: 0.312213

* Train accuracy / confusion: 87.88% / [[324, 20, 11], [29, 225, 15], [7, 15, 154]], 
* Val   accuracy / confusion: 53.85% / [[25, 17, 4], [7, 18, 10], [3, 7, 13]]

------------------------------ Epoch 403 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.618811
- Iter 024 / 025, Loss: 0.349873

* Train accuracy / confusion: 86.88% / [[325, 21, 9], [29, 224, 18], [15, 13, 146]], 
* Val   accuracy / confusion: 46.15% / [[26, 13, 7], [12, 14, 9], [6, 9, 8]]

------------------------------ Epoch 404 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.287781
- Iter 024 / 025, Loss: 0.257515

* Train accuracy / confusion: 87.50% / [[325, 17, 11], [27, 228, 16], [15, 14, 147]], 
* Val   accuracy / confusion: 59.62% / [[30, 10, 6], [6, 23, 6], [10, 4, 9]]

------------------------------ Epoch 405 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.393757
- Iter 024 / 025, Loss: 0.337567

* Train accuracy / confusion: 86.50% / [[325, 19, 10], [31, 222, 21], [5, 22, 145]], 
* Val   accuracy / confusion: 45.19% / [[30, 9, 7], [15, 10, 10], [6, 10, 7]]

------------------------------ Epoch 406 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.208691
- Iter 024 / 025, Loss: 0.260244

* Train accuracy / confusion: 89.12% / [[323, 25, 6], [26, 237, 7], [11, 12, 153]], 
* Val   accuracy / confusion: 53.85% / [[28, 12, 6], [9, 17, 9], [3, 9, 11]]

------------------------------ Epoch 407 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.287296
- Iter 024 / 025, Loss: 0.600012

* Train accuracy / confusion: 88.38% / [[324, 18, 11], [28, 228, 12], [5, 19, 155]], 
* Val   accuracy / confusion: 52.88% / [[29, 10, 7], [9, 18, 8], [7, 8, 8]]

------------------------------ Epoch 408 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.369109
- Iter 024 / 025, Loss: 0.462375

* Train accuracy / confusion: 89.38% / [[332, 17, 6], [26, 228, 14], [12, 10, 155]], 
* Val   accuracy / confusion: 55.77% / [[25, 17, 4], [7, 20, 8], [2, 8, 13]]

------------------------------ Epoch 409 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.274961
- Iter 024 / 025, Loss: 0.289261

* Train accuracy / confusion: 90.88% / [[340, 13, 5], [23, 233, 14], [9, 9, 154]], 
* Val   accuracy / confusion: 57.69% / [[31, 8, 7], [10, 18, 7], [7, 5, 11]]

------------------------------ Epoch 410 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.201039
- Iter 024 / 025, Loss: 0.338663

* Train accuracy / confusion: 88.88% / [[334, 16, 4], [26, 224, 19], [11, 13, 153]], 
* Val   accuracy / confusion: 52.88% / [[28, 11, 7], [10, 15, 10], [4, 7, 12]]

------------------------------ Epoch 411 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.244577
- Iter 024 / 025, Loss: 0.296576

* Train accuracy / confusion: 87.00% / [[318, 20, 13], [28, 226, 19], [10, 14, 152]], 
* Val   accuracy / confusion: 50.00% / [[32, 10, 4], [14, 12, 9], [9, 6, 8]]

------------------------------ Epoch 412 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.363473
- Iter 024 / 025, Loss: 0.291815

* Train accuracy / confusion: 86.12% / [[319, 26, 13], [29, 219, 19], [10, 14, 151]], 
* Val   accuracy / confusion: 49.04% / [[29, 11, 6], [10, 15, 10], [7, 9, 7]]

------------------------------ Epoch 413 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.307159
- Iter 024 / 025, Loss: 0.145725

* Train accuracy / confusion: 88.50% / [[342, 14, 4], [28, 222, 17], [13, 16, 144]], 
* Val   accuracy / confusion: 52.88% / [[34, 8, 4], [17, 14, 4], [8, 8, 7]]

------------------------------ Epoch 414 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.193978
- Iter 024 / 025, Loss: 0.406391

* Train accuracy / confusion: 88.88% / [[327, 22, 5], [28, 230, 11], [7, 16, 154]], 
* Val   accuracy / confusion: 50.96% / [[30, 10, 6], [14, 13, 8], [5, 8, 10]]

------------------------------ Epoch 415 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.371845
- Iter 024 / 025, Loss: 0.296265

* Train accuracy / confusion: 87.75% / [[328, 24, 7], [27, 228, 12], [13, 15, 146]], 
* Val   accuracy / confusion: 54.81% / [[29, 11, 6], [14, 17, 4], [4, 8, 11]]

------------------------------ Epoch 416 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.228029
- Iter 024 / 025, Loss: 0.240129

* Train accuracy / confusion: 87.88% / [[330, 19, 6], [24, 227, 19], [15, 14, 146]], 
* Val   accuracy / confusion: 53.85% / [[27, 9, 10], [8, 18, 9], [2, 10, 11]]

------------------------------ Epoch 417 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.181747
- Iter 024 / 025, Loss: 0.170391

* Train accuracy / confusion: 89.00% / [[336, 14, 8], [28, 226, 12], [13, 13, 150]], 
* Val   accuracy / confusion: 50.96% / [[28, 11, 7], [10, 16, 9], [5, 9, 9]]

------------------------------ Epoch 418 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.133430
- Iter 024 / 025, Loss: 0.486841

* Train accuracy / confusion: 88.00% / [[324, 23, 7], [21, 233, 16], [14, 15, 147]], 
* Val   accuracy / confusion: 48.08% / [[21, 18, 7], [11, 16, 8], [5, 5, 13]]

------------------------------ Epoch 419 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.252586
- Iter 024 / 025, Loss: 0.348982

* Train accuracy / confusion: 86.12% / [[323, 29, 7], [31, 221, 16], [10, 18, 145]], 
* Val   accuracy / confusion: 49.04% / [[27, 13, 6], [7, 19, 9], [11, 7, 5]]

------------------------------ Epoch 420 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.334001
- Iter 024 / 025, Loss: 0.224707

* Train accuracy / confusion: 86.50% / [[329, 22, 9], [29, 223, 18], [15, 15, 140]], 
* Val   accuracy / confusion: 55.77% / [[27, 9, 10], [12, 17, 6], [5, 4, 14]]

------------------------------ Epoch 421 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.382628
- Iter 024 / 025, Loss: 0.252269

* Train accuracy / confusion: 89.00% / [[332, 14, 9], [28, 229, 15], [14, 8, 151]], 
* Val   accuracy / confusion: 47.12% / [[26, 11, 9], [11, 16, 8], [6, 10, 7]]

------------------------------ Epoch 422 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.235498
- Iter 024 / 025, Loss: 0.360837

* Train accuracy / confusion: 88.38% / [[329, 23, 4], [25, 225, 18], [9, 14, 153]], 
* Val   accuracy / confusion: 50.00% / [[25, 13, 8], [13, 17, 5], [5, 8, 10]]

------------------------------ Epoch 423 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.201309
- Iter 024 / 025, Loss: 0.161189

* Train accuracy / confusion: 88.38% / [[332, 19, 5], [27, 226, 13], [15, 14, 149]], 
* Val   accuracy / confusion: 49.04% / [[27, 12, 7], [13, 19, 3], [8, 10, 5]]

------------------------------ Epoch 424 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.336204
- Iter 024 / 025, Loss: 0.400265

* Train accuracy / confusion: 87.25% / [[317, 26, 12], [27, 228, 15], [11, 11, 153]], 
* Val   accuracy / confusion: 54.81% / [[30, 11, 5], [9, 16, 10], [5, 7, 11]]

------------------------------ Epoch 425 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.373840
- Iter 024 / 025, Loss: 0.550738

* Train accuracy / confusion: 87.38% / [[328, 23, 6], [31, 219, 20], [5, 16, 152]], 
* Val   accuracy / confusion: 49.04% / [[29, 13, 4], [12, 15, 8], [4, 12, 7]]

------------------------------ Epoch 426 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.196299
- Iter 024 / 025, Loss: 0.164198

* Train accuracy / confusion: 87.25% / [[335, 18, 5], [30, 212, 24], [12, 13, 151]], 
* Val   accuracy / confusion: 50.96% / [[29, 12, 5], [11, 15, 9], [5, 9, 9]]

------------------------------ Epoch 427 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.216729
- Iter 024 / 025, Loss: 0.341988

* Train accuracy / confusion: 88.38% / [[324, 23, 11], [25, 233, 8], [12, 14, 150]], 
* Val   accuracy / confusion: 50.00% / [[26, 10, 10], [9, 17, 9], [5, 9, 9]]

------------------------------ Epoch 428 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.187450
- Iter 024 / 025, Loss: 0.160753

* Train accuracy / confusion: 86.50% / [[323, 21, 9], [29, 224, 16], [14, 19, 145]], 
* Val   accuracy / confusion: 48.08% / [[28, 12, 6], [10, 15, 10], [5, 11, 7]]

------------------------------ Epoch 429 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.235862
- Iter 024 / 025, Loss: 0.466799

* Train accuracy / confusion: 86.25% / [[331, 26, 6], [31, 213, 22], [14, 11, 146]], 
* Val   accuracy / confusion: 48.08% / [[29, 9, 8], [16, 10, 9], [6, 6, 11]]

------------------------------ Epoch 430 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.601160
- Iter 024 / 025, Loss: 0.239012

* Train accuracy / confusion: 87.12% / [[328, 19, 7], [32, 221, 15], [10, 20, 148]], 
* Val   accuracy / confusion: 44.23% / [[26, 15, 5], [12, 14, 9], [6, 11, 6]]

------------------------------ Epoch 431 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.301855
- Iter 024 / 025, Loss: 0.163165

* Train accuracy / confusion: 89.25% / [[338, 13, 5], [25, 226, 17], [10, 16, 150]], 
* Val   accuracy / confusion: 50.96% / [[28, 12, 6], [10, 17, 8], [8, 7, 8]]

------------------------------ Epoch 432 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.279598
- Iter 024 / 025, Loss: 0.460054

* Train accuracy / confusion: 87.62% / [[335, 12, 10], [28, 227, 19], [13, 17, 139]], 
* Val   accuracy / confusion: 44.23% / [[28, 6, 12], [12, 8, 15], [8, 5, 10]]

------------------------------ Epoch 433 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.515698
- Iter 024 / 025, Loss: 0.360119

* Train accuracy / confusion: 87.38% / [[318, 21, 12], [28, 230, 14], [11, 15, 151]], 
* Val   accuracy / confusion: 51.92% / [[28, 14, 4], [12, 17, 6], [7, 7, 9]]

------------------------------ Epoch 434 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.295607
- Iter 024 / 025, Loss: 0.288127

* Train accuracy / confusion: 87.62% / [[328, 21, 7], [31, 224, 15], [5, 20, 149]], 
* Val   accuracy / confusion: 52.88% / [[26, 13, 7], [12, 16, 7], [3, 7, 13]]

------------------------------ Epoch 435 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.231176
- Iter 024 / 025, Loss: 0.143783

* Train accuracy / confusion: 88.88% / [[326, 24, 5], [31, 231, 9], [9, 11, 154]], 
* Val   accuracy / confusion: 52.88% / [[27, 14, 5], [9, 15, 11], [4, 6, 13]]

------------------------------ Epoch 436 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.273047
- Iter 024 / 025, Loss: 0.265006

* Train accuracy / confusion: 88.38% / [[332, 16, 10], [24, 226, 15], [6, 22, 149]], 
* Val   accuracy / confusion: 50.96% / [[28, 12, 6], [10, 17, 8], [8, 7, 8]]

------------------------------ Epoch 437 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.174584
- Iter 024 / 025, Loss: 0.101566

* Train accuracy / confusion: 88.38% / [[334, 15, 5], [30, 224, 14], [15, 14, 149]], 
* Val   accuracy / confusion: 47.12% / [[27, 13, 6], [11, 17, 7], [9, 9, 5]]

------------------------------ Epoch 438 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.215371
- Iter 024 / 025, Loss: 0.360082

* Train accuracy / confusion: 85.50% / [[326, 24, 7], [30, 218, 18], [16, 21, 140]], 
* Val   accuracy / confusion: 50.96% / [[25, 14, 7], [9, 20, 6], [5, 10, 8]]

------------------------------ Epoch 439 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.287273
- Iter 024 / 025, Loss: 0.260042

* Train accuracy / confusion: 88.50% / [[326, 21, 11], [22, 227, 14], [9, 15, 155]], 
* Val   accuracy / confusion: 54.81% / [[29, 8, 9], [9, 17, 9], [5, 7, 11]]

------------------------------ Epoch 440 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.359112
- Iter 024 / 025, Loss: 0.314690

* Train accuracy / confusion: 86.00% / [[321, 19, 17], [27, 224, 17], [14, 18, 143]], 
* Val   accuracy / confusion: 46.15% / [[27, 13, 6], [14, 14, 7], [6, 10, 7]]

------------------------------ Epoch 441 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.198307
- Iter 024 / 025, Loss: 0.405440

* Train accuracy / confusion: 85.88% / [[324, 24, 10], [41, 214, 12], [7, 19, 149]], 
* Val   accuracy / confusion: 50.00% / [[29, 8, 9], [10, 16, 9], [11, 5, 7]]

------------------------------ Epoch 442 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.320772
- Iter 024 / 025, Loss: 0.556042

* Train accuracy / confusion: 86.25% / [[324, 23, 9], [36, 217, 15], [13, 14, 149]], 
* Val   accuracy / confusion: 52.88% / [[31, 7, 8], [10, 16, 9], [8, 7, 8]]

------------------------------ Epoch 443 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.319970
- Iter 024 / 025, Loss: 0.231874

* Train accuracy / confusion: 87.25% / [[325, 24, 14], [25, 224, 14], [11, 14, 149]], 
* Val   accuracy / confusion: 48.08% / [[26, 15, 5], [10, 13, 12], [6, 6, 11]]

------------------------------ Epoch 444 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.220361
- Iter 024 / 025, Loss: 0.204234

* Train accuracy / confusion: 88.00% / [[326, 26, 2], [31, 221, 18], [6, 13, 157]], 
* Val   accuracy / confusion: 50.00% / [[26, 13, 7], [7, 20, 8], [7, 10, 6]]

------------------------------ Epoch 445 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.315911
- Iter 024 / 025, Loss: 0.152689

* Train accuracy / confusion: 89.25% / [[332, 18, 8], [27, 226, 13], [10, 10, 156]], 
* Val   accuracy / confusion: 50.96% / [[25, 13, 8], [12, 16, 7], [5, 6, 12]]

------------------------------ Epoch 446 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.244980
- Iter 024 / 025, Loss: 0.224498

* Train accuracy / confusion: 89.75% / [[335, 15, 5], [22, 236, 15], [7, 18, 147]], 
* Val   accuracy / confusion: 48.08% / [[31, 11, 4], [15, 13, 7], [8, 9, 6]]

------------------------------ Epoch 447 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.290472
- Iter 024 / 025, Loss: 0.193631

* Train accuracy / confusion: 88.88% / [[326, 22, 6], [27, 233, 11], [9, 14, 152]], 
* Val   accuracy / confusion: 47.12% / [[25, 14, 7], [13, 15, 7], [5, 9, 9]]

------------------------------ Epoch 448 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.342161
- Iter 024 / 025, Loss: 0.223108

* Train accuracy / confusion: 88.75% / [[332, 20, 8], [29, 227, 11], [12, 10, 151]], 
* Val   accuracy / confusion: 52.88% / [[29, 13, 4], [12, 15, 8], [5, 7, 11]]

------------------------------ Epoch 449 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.298332
- Iter 024 / 025, Loss: 0.186762

* Train accuracy / confusion: 87.38% / [[324, 23, 6], [24, 231, 14], [17, 17, 144]], 
* Val   accuracy / confusion: 52.88% / [[30, 12, 4], [11, 16, 8], [5, 9, 9]]

------------------------------ Epoch 450 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.190221
- Iter 024 / 025, Loss: 0.338050

* Train accuracy / confusion: 87.62% / [[336, 14, 9], [29, 219, 20], [6, 21, 146]], 
* Val   accuracy / confusion: 55.77% / [[29, 8, 9], [7, 20, 8], [7, 7, 9]]

------------------------------ Epoch 451 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.210837
- Iter 024 / 025, Loss: 0.262231

* Train accuracy / confusion: 87.00% / [[330, 18, 6], [29, 223, 17], [15, 19, 143]], 
* Val   accuracy / confusion: 50.00% / [[28, 11, 7], [11, 16, 8], [9, 6, 8]]

------------------------------ Epoch 452 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.156503
- Iter 024 / 025, Loss: 0.235861

* Train accuracy / confusion: 89.38% / [[326, 18, 5], [31, 233, 9], [12, 10, 156]], 
* Val   accuracy / confusion: 51.92% / [[27, 11, 8], [8, 19, 8], [7, 8, 8]]

------------------------------ Epoch 453 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.123134
- Iter 024 / 025, Loss: 0.408443

* Train accuracy / confusion: 89.75% / [[331, 13, 10], [29, 231, 9], [8, 13, 156]], 
* Val   accuracy / confusion: 51.92% / [[30, 9, 7], [15, 16, 4], [8, 7, 8]]

------------------------------ Epoch 454 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.210202
- Iter 024 / 025, Loss: 0.395897

* Train accuracy / confusion: 87.75% / [[332, 16, 6], [27, 220, 23], [14, 12, 150]], 
* Val   accuracy / confusion: 53.85% / [[26, 13, 7], [8, 15, 12], [5, 3, 15]]

------------------------------ Epoch 455 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.385142
- Iter 024 / 025, Loss: 0.214778

* Train accuracy / confusion: 87.75% / [[329, 16, 10], [28, 228, 14], [9, 21, 145]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [11, 16, 8], [7, 7, 9]]

------------------------------ Epoch 456 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.284608
- Iter 024 / 025, Loss: 0.275861

* Train accuracy / confusion: 90.12% / [[333, 16, 4], [25, 227, 17], [10, 7, 161]], 
* Val   accuracy / confusion: 53.85% / [[26, 11, 9], [7, 18, 10], [2, 9, 12]]

------------------------------ Epoch 457 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.392026
- Iter 024 / 025, Loss: 0.216346

* Train accuracy / confusion: 89.62% / [[328, 19, 5], [26, 227, 17], [8, 8, 162]], 
* Val   accuracy / confusion: 46.15% / [[22, 18, 6], [9, 15, 11], [5, 7, 11]]

------------------------------ Epoch 458 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.328952
- Iter 024 / 025, Loss: 0.319175

* Train accuracy / confusion: 86.00% / [[325, 21, 13], [27, 221, 18], [15, 18, 142]], 
* Val   accuracy / confusion: 47.12% / [[27, 14, 5], [17, 13, 5], [8, 6, 9]]

------------------------------ Epoch 459 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.167688
- Iter 024 / 025, Loss: 0.297931

* Train accuracy / confusion: 87.88% / [[327, 17, 6], [36, 225, 12], [10, 16, 151]], 
* Val   accuracy / confusion: 53.85% / [[30, 8, 8], [9, 17, 9], [5, 9, 9]]

------------------------------ Epoch 460 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.303091
- Iter 024 / 025, Loss: 0.344629

* Train accuracy / confusion: 88.38% / [[329, 15, 11], [28, 229, 13], [13, 13, 149]], 
* Val   accuracy / confusion: 57.69% / [[34, 7, 5], [13, 13, 9], [3, 7, 13]]

------------------------------ Epoch 461 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.318328
- Iter 024 / 025, Loss: 0.540706

* Train accuracy / confusion: 89.50% / [[332, 14, 9], [17, 236, 16], [17, 11, 148]], 
* Val   accuracy / confusion: 51.92% / [[29, 8, 9], [11, 15, 9], [7, 6, 10]]

------------------------------ Epoch 462 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.185170
- Iter 024 / 025, Loss: 0.685045

* Train accuracy / confusion: 87.75% / [[330, 19, 5], [28, 218, 24], [12, 10, 154]], 
* Val   accuracy / confusion: 50.00% / [[23, 17, 6], [10, 18, 7], [5, 7, 11]]

------------------------------ Epoch 463 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.192257
- Iter 024 / 025, Loss: 0.398548

* Train accuracy / confusion: 89.75% / [[336, 19, 3], [21, 234, 14], [9, 16, 148]], 
* Val   accuracy / confusion: 50.96% / [[30, 11, 5], [13, 16, 6], [5, 11, 7]]

------------------------------ Epoch 464 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.293975
- Iter 024 / 025, Loss: 0.153677

* Train accuracy / confusion: 87.75% / [[328, 20, 10], [33, 224, 7], [10, 18, 150]], 
* Val   accuracy / confusion: 52.88% / [[29, 10, 7], [7, 18, 10], [5, 10, 8]]

------------------------------ Epoch 465 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.209456
- Iter 024 / 025, Loss: 0.328569

* Train accuracy / confusion: 87.62% / [[324, 20, 10], [27, 229, 18], [13, 11, 148]], 
* Val   accuracy / confusion: 51.92% / [[28, 12, 6], [12, 15, 8], [11, 1, 11]]

------------------------------ Epoch 466 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.235364
- Iter 024 / 025, Loss: 0.106603

* Train accuracy / confusion: 89.88% / [[328, 17, 6], [28, 235, 11], [8, 11, 156]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [12, 14, 9], [8, 5, 10]]

------------------------------ Epoch 467 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.180390
- Iter 024 / 025, Loss: 0.093428

* Train accuracy / confusion: 90.38% / [[331, 23, 7], [18, 237, 12], [8, 9, 155]], 
* Val   accuracy / confusion: 50.96% / [[26, 14, 6], [9, 17, 9], [5, 8, 10]]

------------------------------ Epoch 468 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.384287
- Iter 024 / 025, Loss: 0.400452

* Train accuracy / confusion: 88.38% / [[334, 20, 6], [31, 221, 14], [6, 16, 152]], 
* Val   accuracy / confusion: 52.88% / [[29, 9, 8], [10, 19, 6], [6, 10, 7]]

------------------------------ Epoch 469 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.395957
- Iter 024 / 025, Loss: 0.434298

* Train accuracy / confusion: 87.38% / [[326, 19, 10], [30, 225, 14], [9, 19, 148]], 
* Val   accuracy / confusion: 49.04% / [[27, 12, 7], [13, 14, 8], [9, 4, 10]]

------------------------------ Epoch 470 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.416960
- Iter 024 / 025, Loss: 0.508894

* Train accuracy / confusion: 85.25% / [[322, 28, 7], [33, 220, 13], [11, 26, 140]], 
* Val   accuracy / confusion: 52.88% / [[25, 11, 10], [8, 22, 5], [2, 13, 8]]

------------------------------ Epoch 471 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.115569
- Iter 024 / 025, Loss: 0.452710

* Train accuracy / confusion: 89.12% / [[331, 15, 6], [24, 229, 17], [9, 16, 153]], 
* Val   accuracy / confusion: 45.19% / [[28, 13, 5], [13, 13, 9], [8, 9, 6]]

------------------------------ Epoch 472 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.429078
- Iter 024 / 025, Loss: 0.272587

* Train accuracy / confusion: 86.88% / [[327, 20, 5], [30, 223, 18], [10, 22, 145]], 
* Val   accuracy / confusion: 53.85% / [[27, 13, 6], [10, 16, 9], [4, 6, 13]]

------------------------------ Epoch 473 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.253132
- Iter 024 / 025, Loss: 0.546936

* Train accuracy / confusion: 86.62% / [[324, 22, 9], [26, 219, 19], [16, 15, 150]], 
* Val   accuracy / confusion: 55.77% / [[25, 12, 9], [10, 20, 5], [3, 7, 13]]

------------------------------ Epoch 474 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.208681
- Iter 024 / 025, Loss: 0.295108

* Train accuracy / confusion: 87.88% / [[333, 16, 7], [24, 230, 15], [14, 21, 140]], 
* Val   accuracy / confusion: 48.08% / [[27, 12, 7], [13, 15, 7], [9, 6, 8]]

------------------------------ Epoch 475 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.140888
- Iter 024 / 025, Loss: 0.354460

* Train accuracy / confusion: 91.12% / [[331, 16, 8], [25, 236, 5], [7, 10, 162]], 
* Val   accuracy / confusion: 45.19% / [[26, 13, 7], [13, 13, 9], [11, 4, 8]]

------------------------------ Epoch 476 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.434271
- Iter 024 / 025, Loss: 0.282860

* Train accuracy / confusion: 89.75% / [[336, 18, 3], [22, 231, 11], [10, 18, 151]], 
* Val   accuracy / confusion: 50.00% / [[26, 10, 10], [9, 16, 10], [5, 8, 10]]

------------------------------ Epoch 477 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.294761
- Iter 024 / 025, Loss: 0.468881

* Train accuracy / confusion: 87.12% / [[321, 29, 10], [28, 226, 12], [11, 13, 150]], 
* Val   accuracy / confusion: 59.62% / [[30, 7, 9], [4, 21, 10], [6, 6, 11]]

------------------------------ Epoch 478 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.340003
- Iter 024 / 025, Loss: 0.483693

* Train accuracy / confusion: 86.75% / [[326, 21, 11], [35, 219, 13], [10, 16, 149]], 
* Val   accuracy / confusion: 58.65% / [[30, 11, 5], [8, 17, 10], [2, 7, 14]]

------------------------------ Epoch 479 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.602817
- Iter 024 / 025, Loss: 0.335670

* Train accuracy / confusion: 87.62% / [[331, 16, 7], [31, 219, 18], [9, 18, 151]], 
* Val   accuracy / confusion: 47.12% / [[28, 12, 6], [12, 15, 8], [6, 11, 6]]

------------------------------ Epoch 480 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.190479
- Iter 024 / 025, Loss: 0.258974

* Train accuracy / confusion: 87.75% / [[329, 19, 9], [27, 221, 22], [9, 12, 152]], 
* Val   accuracy / confusion: 53.85% / [[31, 10, 5], [12, 16, 7], [8, 6, 9]]

------------------------------ Epoch 481 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.142359
- Iter 024 / 025, Loss: 0.427703

* Train accuracy / confusion: 88.00% / [[325, 24, 6], [27, 228, 17], [7, 15, 151]], 
* Val   accuracy / confusion: 50.00% / [[26, 10, 10], [15, 15, 5], [5, 7, 11]]

------------------------------ Epoch 482 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.300744
- Iter 024 / 025, Loss: 0.293114

* Train accuracy / confusion: 87.88% / [[326, 20, 7], [34, 228, 9], [11, 16, 149]], 
* Val   accuracy / confusion: 52.88% / [[28, 8, 10], [11, 18, 6], [3, 11, 9]]

------------------------------ Epoch 483 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.290219
- Iter 024 / 025, Loss: 0.300592

* Train accuracy / confusion: 90.12% / [[339, 8, 8], [32, 223, 13], [7, 11, 159]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [9, 20, 6], [10, 8, 5]]

------------------------------ Epoch 484 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.224382
- Iter 024 / 025, Loss: 0.210399

* Train accuracy / confusion: 87.25% / [[317, 27, 8], [36, 222, 13], [9, 9, 159]], 
* Val   accuracy / confusion: 50.96% / [[30, 9, 7], [11, 17, 7], [7, 10, 6]]

------------------------------ Epoch 485 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.170867
- Iter 024 / 025, Loss: 0.549963

* Train accuracy / confusion: 88.50% / [[330, 18, 4], [33, 220, 15], [13, 9, 158]], 
* Val   accuracy / confusion: 50.96% / [[31, 10, 5], [14, 14, 7], [6, 9, 8]]

------------------------------ Epoch 486 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.167517
- Iter 024 / 025, Loss: 0.267441

* Train accuracy / confusion: 89.25% / [[338, 13, 5], [23, 236, 13], [13, 19, 140]], 
* Val   accuracy / confusion: 52.88% / [[25, 9, 12], [9, 21, 5], [9, 5, 9]]

------------------------------ Epoch 487 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.320276
- Iter 024 / 025, Loss: 0.150021

* Train accuracy / confusion: 89.62% / [[329, 24, 1], [22, 228, 23], [6, 7, 160]], 
* Val   accuracy / confusion: 49.04% / [[26, 11, 9], [7, 18, 10], [7, 9, 7]]

------------------------------ Epoch 488 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.456934
- Iter 024 / 025, Loss: 0.373984

* Train accuracy / confusion: 88.00% / [[331, 16, 8], [30, 223, 16], [9, 17, 150]], 
* Val   accuracy / confusion: 54.81% / [[34, 7, 5], [12, 16, 7], [5, 11, 7]]

------------------------------ Epoch 489 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.370364
- Iter 024 / 025, Loss: 0.182286

* Train accuracy / confusion: 87.88% / [[323, 22, 14], [30, 224, 10], [8, 13, 156]], 
* Val   accuracy / confusion: 51.92% / [[31, 8, 7], [11, 15, 9], [9, 6, 8]]

------------------------------ Epoch 490 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.331486
- Iter 024 / 025, Loss: 0.376345

* Train accuracy / confusion: 88.00% / [[327, 21, 7], [33, 225, 14], [8, 13, 152]], 
* Val   accuracy / confusion: 52.88% / [[31, 8, 7], [14, 17, 4], [9, 7, 7]]

------------------------------ Epoch 491 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.159715
- Iter 024 / 025, Loss: 0.328947

* Train accuracy / confusion: 89.25% / [[334, 14, 7], [22, 232, 16], [12, 15, 148]], 
* Val   accuracy / confusion: 50.96% / [[27, 13, 6], [13, 17, 5], [6, 8, 9]]

------------------------------ Epoch 492 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.155318
- Iter 024 / 025, Loss: 0.219722

* Train accuracy / confusion: 90.12% / [[336, 14, 8], [23, 234, 9], [10, 15, 151]], 
* Val   accuracy / confusion: 55.77% / [[31, 9, 6], [10, 18, 7], [7, 7, 9]]

------------------------------ Epoch 493 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.303653
- Iter 024 / 025, Loss: 0.506403

* Train accuracy / confusion: 87.75% / [[326, 19, 10], [27, 224, 18], [11, 13, 152]], 
* Val   accuracy / confusion: 52.88% / [[31, 8, 7], [14, 15, 6], [6, 8, 9]]

------------------------------ Epoch 494 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.349390
- Iter 024 / 025, Loss: 0.467372

* Train accuracy / confusion: 89.00% / [[337, 15, 6], [29, 228, 11], [16, 11, 147]], 
* Val   accuracy / confusion: 49.04% / [[26, 12, 8], [11, 16, 8], [4, 10, 9]]

------------------------------ Epoch 495 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.143137
- Iter 024 / 025, Loss: 0.315900

* Train accuracy / confusion: 87.25% / [[329, 21, 5], [25, 224, 19], [16, 16, 145]], 
* Val   accuracy / confusion: 52.88% / [[27, 14, 5], [8, 19, 8], [6, 8, 9]]

------------------------------ Epoch 496 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.205844
- Iter 024 / 025, Loss: 0.144735

* Train accuracy / confusion: 89.00% / [[337, 11, 6], [23, 227, 21], [11, 16, 148]], 
* Val   accuracy / confusion: 48.08% / [[26, 12, 8], [12, 15, 8], [8, 6, 9]]

------------------------------ Epoch 497 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.357824
- Iter 024 / 025, Loss: 0.285507

* Train accuracy / confusion: 87.88% / [[329, 21, 9], [27, 230, 13], [10, 17, 144]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [8, 16, 11], [5, 10, 8]]

------------------------------ Epoch 498 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.373563
- Iter 024 / 025, Loss: 0.188903

* Train accuracy / confusion: 89.38% / [[328, 22, 7], [20, 235, 15], [7, 14, 152]], 
* Val   accuracy / confusion: 50.96% / [[29, 10, 7], [11, 16, 8], [8, 7, 8]]

------------------------------ Epoch 499 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.371795
- Iter 024 / 025, Loss: 0.161637

* Train accuracy / confusion: 88.75% / [[330, 13, 11], [28, 230, 12], [11, 15, 150]], 
* Val   accuracy / confusion: 54.81% / [[28, 11, 7], [11, 17, 7], [4, 7, 12]]

------------------------------ Epoch 500 / 500, Learning rate: 3.81e-05 ------------------------------
- Iter 012 / 025, Loss: 0.175311
- Iter 024 / 025, Loss: 0.202215

* Train accuracy / confusion: 88.75% / [[339, 15, 3], [27, 227, 15], [10, 20, 144]], 
* Val   accuracy / confusion: 50.00% / [[27, 11, 8], [10, 17, 8], [6, 9, 8]]

**************************************** Training Ends ****************************************
- Test accuracy: 51.62%

- Confusion matrix:
 [[862 397 121]
 [258 420 342]
 [102 275 313]]

- Debug table:
{   '00941': {'GT': 0, 'Acc': ' 46.67%', 'Pred': [14, 16, 0]},
    '00408': {'GT': 0, 'Acc': ' 30.00%', 'Pred': [9, 17, 4]},
    '00871': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00852': {'GT': 0, 'Acc': ' 10.00%', 'Pred': [3, 27, 0]},
    '00816': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00672': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00414': {'GT': 2, 'Acc': ' 30.00%', 'Pred': [5, 16, 9]},
    '00767': {'GT': 1, 'Acc': ' 30.00%', 'Pred': [20, 9, 1]},
    '01148': {'GT': 1, 'Acc': ' 13.33%', 'Pred': [4, 4, 22]},
    '00743': {'GT': 0, 'Acc': ' 33.33%', 'Pred': [10, 10, 10]},
    '00305': {'GT': 1, 'Acc': ' 36.67%', 'Pred': [5, 11, 14]},
    '01305': {'GT': 1, 'Acc': ' 80.00%', 'Pred': [6, 24, 0]},
    '00455': {'GT': 1, 'Acc': '  6.67%', 'Pred': [4, 2, 24]},
    '01140': {'GT': 0, 'Acc': ' 83.33%', 'Pred': [25, 5, 0]},
    '00649': {'GT': 2, 'Acc': ' 70.00%', 'Pred': [6, 3, 21]},
    '01068': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00176': {'GT': 0, 'Acc': ' 26.67%', 'Pred': [8, 22, 0]},
    '00861': {'GT': 2, 'Acc': ' 33.33%', 'Pred': [5, 15, 10]},
    '00302': {'GT': 1, 'Acc': '  0.00%', 'Pred': [5, 0, 25]},
    '00756': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00706': {'GT': 0, 'Acc': ' 80.00%', 'Pred': [24, 6, 0]},
    '01218': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 21, 9]},
    '01008': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00914': {'GT': 0, 'Acc': ' 56.67%', 'Pred': [17, 13, 0]},
    '00680': {'GT': 1, 'Acc': ' 93.33%', 'Pred': [0, 28, 2]},
    '00736': {'GT': 2, 'Acc': ' 66.67%', 'Pred': [0, 10, 20]},
    '00417': {'GT': 2, 'Acc': ' 66.67%', 'Pred': [4, 6, 20]},
    '01184': {'GT': 2, 'Acc': ' 70.00%', 'Pred': [9, 0, 21]},
    '00949': {'GT': 1, 'Acc': ' 86.67%', 'Pred': [1, 26, 3]},
    '01268': {'GT': 1, 'Acc': ' 60.00%', 'Pred': [12, 18, 0]},
    '00441': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00731': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00806': {'GT': 2, 'Acc': ' 16.67%', 'Pred': [13, 12, 5]},
    '00192': {'GT': 0, 'Acc': ' 46.67%', 'Pred': [14, 14, 2]},
    '01075': {'GT': 2, 'Acc': ' 26.67%', 'Pred': [0, 22, 8]},
    '00058': {'GT': 0, 'Acc': ' 76.67%', 'Pred': [23, 1, 6]},
    '01126': {'GT': 0, 'Acc': ' 93.33%', 'Pred': [28, 2, 0]},
    '00924': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01070': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00901': {'GT': 0, 'Acc': '  6.67%', 'Pred': [2, 28, 0]},
    '01079': {'GT': 1, 'Acc': '  6.67%', 'Pred': [25, 2, 3]},
    '00614': {'GT': 0, 'Acc': ' 93.33%', 'Pred': [28, 0, 2]},
    '00062': {'GT': 1, 'Acc': ' 86.67%', 'Pred': [0, 26, 4]},
    '00405': {'GT': 2, 'Acc': ' 80.00%', 'Pred': [3, 3, 24]},
    '01307': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00918': {'GT': 0, 'Acc': '  6.67%', 'Pred': [2, 11, 17]},
    '00299': {'GT': 0, 'Acc': ' 13.33%', 'Pred': [4, 9, 17]},
    '01046': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01106': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00822': {'GT': 0, 'Acc': ' 83.33%', 'Pred': [25, 0, 5]},
    '01037': {'GT': 1, 'Acc': ' 20.00%', 'Pred': [0, 6, 24]},
    '00599': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00517': {'GT': 0, 'Acc': ' 60.00%', 'Pred': [18, 12, 0]},
    '00832': {'GT': 2, 'Acc': ' 16.67%', 'Pred': [7, 18, 5]},
    '00498': {'GT': 1, 'Acc': '  0.00%', 'Pred': [26, 0, 4]},
    '00280': {'GT': 2, 'Acc': '  6.67%', 'Pred': [1, 27, 2]},
    '00623': {'GT': 2, 'Acc': ' 83.33%', 'Pred': [0, 5, 25]},
    '01038': {'GT': 1, 'Acc': ' 13.33%', 'Pred': [0, 4, 26]},
    '00534': {'GT': 0, 'Acc': ' 20.00%', 'Pred': [6, 24, 0]},
    '01199': {'GT': 2, 'Acc': ' 43.33%', 'Pred': [5, 12, 13]},
    '00886': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00698': {'GT': 1, 'Acc': '  6.67%', 'Pred': [0, 2, 28]},
    '00509': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 4, 26]},
    '00389': {'GT': 1, 'Acc': ' 70.00%', 'Pred': [2, 21, 7]},
    '01208': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00004': {'GT': 2, 'Acc': ' 86.67%', 'Pred': [0, 4, 26]},
    '01039': {'GT': 1, 'Acc': ' 76.67%', 'Pred': [0, 23, 7]},
    '01086': {'GT': 2, 'Acc': ' 43.33%', 'Pred': [7, 10, 13]},
    '00174': {'GT': 2, 'Acc': ' 13.33%', 'Pred': [26, 0, 4]},
    '00226': {'GT': 2, 'Acc': ' 86.67%', 'Pred': [0, 4, 26]},
    '00794': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00289': {'GT': 2, 'Acc': ' 56.67%', 'Pred': [4, 9, 17]},
    '01027': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 3, 0]},
    '00855': {'GT': 0, 'Acc': ' 70.00%', 'Pred': [21, 9, 0]},
    '00385': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 12, 18]},
    '00022': {'GT': 1, 'Acc': ' 16.67%', 'Pred': [13, 5, 12]},
    '00220': {'GT': 1, 'Acc': ' 10.00%', 'Pred': [18, 3, 9]},
    '00635': {'GT': 2, 'Acc': ' 36.67%', 'Pred': [4, 15, 11]},
    '00122': {'GT': 0, 'Acc': ' 60.00%', 'Pred': [18, 12, 0]},
    '00505': {'GT': 0, 'Acc': ' 53.33%', 'Pred': [16, 12, 2]},
    '01014': {'GT': 1, 'Acc': '  3.33%', 'Pred': [29, 1, 0]},
    '00407': {'GT': 1, 'Acc': ' 30.00%', 'Pred': [21, 9, 0]},
    '01300': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00301': {'GT': 1, 'Acc': '  3.33%', 'Pred': [0, 1, 29]},
    '01349': {'GT': 1, 'Acc': ' 96.67%', 'Pred': [1, 29, 0]},
    '00005': {'GT': 2, 'Acc': ' 33.33%', 'Pred': [3, 17, 10]},
    '01169': {'GT': 0, 'Acc': ' 33.33%', 'Pred': [10, 20, 0]},
    '00124': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00243': {'GT': 1, 'Acc': ' 60.00%', 'Pred': [11, 18, 1]},
    '01240': {'GT': 1, 'Acc': '  3.33%', 'Pred': [3, 1, 26]},
    '00589': {'GT': 0, 'Acc': ' 40.00%', 'Pred': [12, 15, 3]},
    '01172': {'GT': 1, 'Acc': '  0.00%', 'Pred': [1, 0, 29]},
    '01156': {'GT': 1, 'Acc': ' 86.67%', 'Pred': [4, 26, 0]},
    '00329': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 3, 0]},
    '01333': {'GT': 0, 'Acc': ' 30.00%', 'Pred': [9, 21, 0]},
    '01364': {'GT': 1, 'Acc': ' 86.67%', 'Pred': [0, 26, 4]},
    '00415': {'GT': 1, 'Acc': ' 10.00%', 'Pred': [27, 3, 0]},
    '00603': {'GT': 2, 'Acc': ' 76.67%', 'Pred': [0, 7, 23]},
    '00096': {'GT': 0, 'Acc': ' 83.33%', 'Pred': [25, 5, 0]},
    '01250': {'GT': 1, 'Acc': ' 16.67%', 'Pred': [16, 5, 9]},
    '00125': {'GT': 0, 'Acc': ' 56.67%', 'Pred': [17, 13, 0]},
    '00447': {'GT': 1, 'Acc': ' 66.67%', 'Pred': [0, 20, 10]},
    '00820': {'GT': 1, 'Acc': ' 23.33%', 'Pred': [4, 7, 19]}}

1D CNN: Siamese model3

class SiameseNet3(nn.Module):
    def __init__(self, n_input=20, n_output=3, n_channel=64,
                 use_age=True, final_pool='average'):
        super().__init__()
        
        if final_pool not in {'average', 'max'}:
            raise ValueError("final_pool must be set to one of ['average', 'max']")
        
        self.use_age = use_age
        
        self.conv1 = nn.Conv1d(n_input, n_channel, kernel_size=25, stride=2)
        self.bn1 = nn.BatchNorm1d(n_channel)
        self.pool1 = nn.MaxPool1d(2)
        
        self.conv2 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn2 = nn.BatchNorm1d(n_channel)
        self.pool2 = nn.MaxPool1d(2)
        
        self.conv3 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn3 = nn.BatchNorm1d(n_channel)
        self.pool3 = nn.MaxPool1d(2)
        
        self.conv4_1 = nn.Conv1d(n_channel, n_channel*2, kernel_size=9)
        self.bn4_1 = nn.BatchNorm1d(n_channel*2)

        self.conv4_2 = nn.Conv1d(n_channel, n_channel*2, kernel_size=9)
        self.bn4_2 = nn.BatchNorm1d(n_channel*2)

        self.conv5_1 = nn.Conv1d(n_channel*2, n_channel*2, kernel_size=7)
        self.bn5_1 = nn.BatchNorm1d(n_channel*2)

        self.conv5_2 = nn.Conv1d(n_channel*2, n_channel*2, kernel_size=7)
        self.bn5_2 = nn.BatchNorm1d(n_channel*2)
        
        if final_pool == 'average':
            self.final_pool = nn.AdaptiveAvgPool1d(1)
        elif final_pool == 'max':
            self.final_pool = nn.AdaptiveMaxPool1d(1)
            
        if self.use_age:        
            self.fc1 = nn.Linear(n_channel*2 + 1, n_channel)
        else:
            self.fc1 = nn.Linear(n_channel*2, n_channel)
            
        self.dropout = nn.Dropout(p=0.2)
        self.bnfc1 = nn.BatchNorm1d(n_channel)

        self.fc2 = nn.Linear(n_channel, n_channel//2)
        self.bnfc2 = nn.BatchNorm1d(n_channel//2)
        self.fc3 = nn.Linear(n_channel//2, n_output)
        
    def reset_weights(self):
        for m in self.modules():
            if hasattr(m, 'reset_parameters'):
                m.reset_parameters()

    def forward(self, x, age):
        # conv-bn-relu-pool 
        x = self.conv1(x)
        x = F.relu(self.bn1(x))
        x = self.pool1(x)

        x = self.conv2(x)
        x = F.relu(self.bn2(x))
        x = self.pool2(x)

        x = self.conv3(x)
        x = F.relu(self.bn3(x))
        x = self.pool3(x)
        
        L = x.shape[2]
        x1 = x[:, :, :L//2]
        x2 = x[:, :, L//2:]
        
        x1 = self.conv4_1(x1)
        x1 = F.relu(self.bn4_1(x1))

        x2 = self.conv4_2(x2)
        x2 = F.relu(self.bn4_2(x2))
        
        x1 = self.conv5_1(x1)
        x1 = F.relu(self.bn5_1(x1))

        x2 = self.conv5_2(x2)
        x2 = F.relu(self.bn5_2(x2))
                
        x1 = self.final_pool(x1).squeeze()
        x2 = self.final_pool(x2).squeeze()

        x = x1 * x2
        x = x.reshape(x.shape[0], -1) # (N, C, 1) -> (N, C)
        if self.use_age:
            x = torch.cat((x, age.reshape(-1, 1)), dim=1)

        # fc-bn-dropout-relu-fc
        x = self.fc1(x)
        x = self.bnfc1(x)
        x = self.dropout(x)
        x = F.relu(x)
        
        x = self.fc2(x)
        x = self.bnfc2(x)
        x = F.relu(x)
        
        x = self.fc3(x)
        
        return x
        # return F.log_softmax(x, dim=1)

model = SiameseNet3(n_input=train_dataset[0]['signal'].shape[0], 
                    n_output=3, 
                    use_age=True, 
                    final_pool='max')

model = model.to(device, dtype=torch.float32)
print(model)
print()

# tensorboard visualization
visualize_network_tensorboard(model, 'SiameseNet_3')

# number of parameters
n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

SiameseNet3(
  (conv1): Conv1d(20, 64, kernel_size=(25,), stride=(2,))
  (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool1): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv3): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn3): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool3): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv4_1): Conv1d(64, 128, kernel_size=(9,), stride=(1,))
  (bn4_1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv4_2): Conv1d(64, 128, kernel_size=(9,), stride=(1,))
  (bn4_2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv5_1): Conv1d(128, 128, kernel_size=(7,), stride=(1,))
  (bn5_1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv5_2): Conv1d(128, 128, kernel_size=(7,), stride=(1,))
  (bn5_2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (final_pool): AdaptiveMaxPool1d(output_size=1)
  (fc1): Linear(in_features=129, out_features=64, bias=True)
  (dropout): Dropout(p=0.2, inplace=False)
  (bnfc1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc2): Linear(in_features=64, out_features=32, bias=True)
  (bnfc2): BatchNorm1d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc3): Linear(in_features=32, out_features=3, bias=True)
)

The Number of parameters of the model: 478,979

record = learning_rate_search(model, 
                              min_log_lr=-5.0, 
                              max_log_lr=-1.0, 
                              trials=500, 
                              epochs=1)
draw_learning_rate_record(record)
best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

# best_log_lr = -2.0523682533344423
print('best_log_lr:', best_log_lr)

best_log_lr: -1.484028967247533

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 200
log_interval = len(train_loader) // 2

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d}, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e} {"-"*30}')
    
    # train
    loss, train_accuracy, train_confusion = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy, val_confusion = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train accuracy / confusion: {train_accuracy:.2f}% / {train_confusion.tolist()}, ')
    print(f'* Val   accuracy / confusion: {val_accuracy:.2f}% / {val_confusion.tolist()}')
    print()
    
# test
test_accuracy, test_confusion, test_debug = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print()
print('- Confusion matrix:\n', test_confusion)
print()
print('- Debug table:')
pprint.pp(test_debug, indent=4)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(test_confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.142780
- Iter 024 / 025, Loss: 0.955923

* Train accuracy / confusion: 45.88% / [[305, 35, 16], [211, 53, 7], [122, 42, 9]], 
* Val   accuracy / confusion: 48.08% / [[40, 5, 1], [26, 7, 2], [17, 3, 3]]

------------------------------ Epoch 002 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.110369
- Iter 024 / 025, Loss: 0.844157

* Train accuracy / confusion: 48.38% / [[235, 118, 2], [118, 148, 4], [63, 108, 4]], 
* Val   accuracy / confusion: 47.12% / [[44, 2, 0], [30, 5, 0], [22, 1, 0]]

------------------------------ Epoch 003 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.025864
- Iter 024 / 025, Loss: 0.924406

* Train accuracy / confusion: 48.38% / [[285, 71, 1], [167, 101, 0], [92, 82, 1]], 
* Val   accuracy / confusion: 54.81% / [[30, 16, 0], [8, 27, 0], [4, 19, 0]]

------------------------------ Epoch 004 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.085619
- Iter 024 / 025, Loss: 0.964395

* Train accuracy / confusion: 47.12% / [[227, 122, 6], [117, 144, 7], [54, 117, 6]], 
* Val   accuracy / confusion: 46.15% / [[43, 3, 0], [29, 5, 1], [21, 2, 0]]

------------------------------ Epoch 005 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.998243
- Iter 024 / 025, Loss: 1.073913

* Train accuracy / confusion: 49.50% / [[283, 55, 21], [153, 91, 23], [71, 81, 22]], 
* Val   accuracy / confusion: 55.77% / [[31, 15, 0], [7, 27, 1], [5, 18, 0]]

------------------------------ Epoch 006 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.910699
- Iter 024 / 025, Loss: 1.058335

* Train accuracy / confusion: 48.75% / [[225, 124, 5], [105, 160, 2], [58, 116, 5]], 
* Val   accuracy / confusion: 53.85% / [[34, 11, 1], [13, 22, 0], [9, 14, 0]]

------------------------------ Epoch 007 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.964244
- Iter 024 / 025, Loss: 0.876827

* Train accuracy / confusion: 50.00% / [[246, 111, 4], [114, 153, 0], [56, 115, 1]], 
* Val   accuracy / confusion: 54.81% / [[33, 13, 0], [11, 24, 0], [7, 16, 0]]

------------------------------ Epoch 008 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.056991
- Iter 024 / 025, Loss: 0.989361

* Train accuracy / confusion: 51.12% / [[243, 100, 8], [106, 156, 9], [55, 113, 10]], 
* Val   accuracy / confusion: 57.69% / [[33, 13, 0], [8, 27, 0], [5, 18, 0]]

------------------------------ Epoch 009 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.986293
- Iter 024 / 025, Loss: 0.888801

* Train accuracy / confusion: 51.62% / [[246, 96, 13], [107, 142, 19], [53, 99, 25]], 
* Val   accuracy / confusion: 56.73% / [[33, 13, 0], [9, 25, 1], [5, 17, 1]]

------------------------------ Epoch 010 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.903623
- Iter 024 / 025, Loss: 0.916887

* Train accuracy / confusion: 52.50% / [[248, 94, 14], [92, 140, 33], [53, 94, 32]], 
* Val   accuracy / confusion: 57.69% / [[33, 13, 0], [8, 27, 0], [7, 16, 0]]

------------------------------ Epoch 011 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.914558
- Iter 024 / 025, Loss: 0.951143

* Train accuracy / confusion: 53.00% / [[244, 102, 8], [102, 143, 24], [52, 88, 37]], 
* Val   accuracy / confusion: 56.73% / [[34, 10, 2], [8, 24, 3], [5, 17, 1]]

------------------------------ Epoch 012 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.047787
- Iter 024 / 025, Loss: 0.907447

* Train accuracy / confusion: 51.88% / [[255, 92, 14], [109, 129, 27], [59, 84, 31]], 
* Val   accuracy / confusion: 53.85% / [[39, 7, 0], [18, 17, 0], [14, 9, 0]]

------------------------------ Epoch 013 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.050360
- Iter 024 / 025, Loss: 1.001728

* Train accuracy / confusion: 54.12% / [[267, 71, 16], [110, 116, 41], [63, 66, 50]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [7, 25, 3], [4, 18, 1]]

------------------------------ Epoch 014 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.989650
- Iter 024 / 025, Loss: 0.989231

* Train accuracy / confusion: 54.38% / [[273, 75, 10], [119, 133, 15], [60, 86, 29]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [6, 25, 4], [4, 18, 1]]

------------------------------ Epoch 015 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.035391
- Iter 024 / 025, Loss: 0.852480

* Train accuracy / confusion: 52.00% / [[251, 92, 16], [103, 128, 32], [61, 80, 37]], 
* Val   accuracy / confusion: 58.65% / [[34, 11, 1], [7, 25, 3], [5, 16, 2]]

------------------------------ Epoch 016 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.023849
- Iter 024 / 025, Loss: 0.881519

* Train accuracy / confusion: 55.25% / [[244, 96, 19], [90, 158, 22], [46, 85, 40]], 
* Val   accuracy / confusion: 55.77% / [[31, 14, 1], [7, 26, 2], [5, 17, 1]]

------------------------------ Epoch 017 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.791787
- Iter 024 / 025, Loss: 1.172579

* Train accuracy / confusion: 55.25% / [[265, 81, 12], [105, 134, 27], [57, 76, 43]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [7, 25, 3], [4, 18, 1]]

------------------------------ Epoch 018 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.055616
- Iter 024 / 025, Loss: 0.933650

* Train accuracy / confusion: 53.62% / [[248, 81, 25], [98, 126, 45], [54, 68, 55]], 
* Val   accuracy / confusion: 58.65% / [[33, 12, 1], [6, 27, 2], [4, 18, 1]]

------------------------------ Epoch 019 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.993898
- Iter 024 / 025, Loss: 0.943563

* Train accuracy / confusion: 54.38% / [[245, 90, 15], [97, 147, 27], [51, 85, 43]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [8, 21, 6], [4, 14, 5]]

------------------------------ Epoch 020 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.907785
- Iter 024 / 025, Loss: 0.979114

* Train accuracy / confusion: 53.62% / [[257, 98, 4], [100, 145, 23], [54, 92, 27]], 
* Val   accuracy / confusion: 54.81% / [[28, 17, 1], [3, 28, 4], [4, 18, 1]]

------------------------------ Epoch 021 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.977767
- Iter 024 / 025, Loss: 1.085915

* Train accuracy / confusion: 51.62% / [[243, 93, 23], [102, 121, 44], [52, 73, 49]], 
* Val   accuracy / confusion: 53.85% / [[36, 9, 1], [14, 19, 2], [10, 12, 1]]

------------------------------ Epoch 022 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.907075
- Iter 024 / 025, Loss: 0.996588

* Train accuracy / confusion: 53.88% / [[251, 103, 2], [98, 153, 17], [57, 92, 27]], 
* Val   accuracy / confusion: 52.88% / [[33, 12, 1], [11, 21, 3], [6, 16, 1]]

------------------------------ Epoch 023 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.872017
- Iter 024 / 025, Loss: 0.788711

* Train accuracy / confusion: 53.88% / [[264, 82, 10], [113, 132, 23], [55, 86, 35]], 
* Val   accuracy / confusion: 53.85% / [[34, 10, 2], [11, 21, 3], [6, 16, 1]]

------------------------------ Epoch 024 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.972744
- Iter 024 / 025, Loss: 0.867219

* Train accuracy / confusion: 54.38% / [[264, 77, 13], [109, 120, 40], [63, 63, 51]], 
* Val   accuracy / confusion: 53.85% / [[33, 13, 0], [11, 22, 2], [6, 16, 1]]

------------------------------ Epoch 025 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.995023
- Iter 024 / 025, Loss: 1.025115

* Train accuracy / confusion: 53.75% / [[268, 83, 7], [110, 127, 28], [54, 88, 35]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [10, 17, 8], [6, 14, 3]]

------------------------------ Epoch 026 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.980920
- Iter 024 / 025, Loss: 0.972456

* Train accuracy / confusion: 53.38% / [[263, 81, 14], [110, 121, 37], [52, 79, 43]], 
* Val   accuracy / confusion: 52.88% / [[27, 16, 3], [3, 27, 5], [4, 18, 1]]

------------------------------ Epoch 027 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.909245
- Iter 024 / 025, Loss: 0.906141

* Train accuracy / confusion: 54.00% / [[243, 110, 6], [86, 155, 24], [44, 98, 34]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 20, 8], [5, 14, 4]]

------------------------------ Epoch 028 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.013233
- Iter 024 / 025, Loss: 1.008444

* Train accuracy / confusion: 53.75% / [[257, 83, 16], [111, 119, 38], [48, 74, 54]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 029 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.840574
- Iter 024 / 025, Loss: 1.057327

* Train accuracy / confusion: 54.50% / [[248, 99, 7], [96, 149, 24], [47, 91, 39]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [7, 21, 7], [4, 15, 4]]

------------------------------ Epoch 030 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.911348
- Iter 024 / 025, Loss: 0.928066

* Train accuracy / confusion: 54.25% / [[251, 89, 14], [96, 135, 38], [52, 77, 48]], 
* Val   accuracy / confusion: 54.81% / [[34, 11, 1], [10, 22, 3], [6, 16, 1]]

------------------------------ Epoch 031 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.961896
- Iter 024 / 025, Loss: 0.932607

* Train accuracy / confusion: 52.62% / [[266, 83, 8], [113, 122, 31], [57, 87, 33]], 
* Val   accuracy / confusion: 50.96% / [[33, 8, 5], [10, 15, 10], [5, 13, 5]]

------------------------------ Epoch 032 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.856897
- Iter 024 / 025, Loss: 0.884987

* Train accuracy / confusion: 55.88% / [[263, 77, 17], [108, 126, 35], [53, 63, 58]], 
* Val   accuracy / confusion: 53.85% / [[34, 12, 0], [12, 21, 2], [6, 16, 1]]

------------------------------ Epoch 033 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.017497
- Iter 024 / 025, Loss: 0.864794

* Train accuracy / confusion: 53.88% / [[251, 98, 7], [102, 147, 21], [51, 90, 33]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [6, 24, 5], [4, 18, 1]]

------------------------------ Epoch 034 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.952792
- Iter 024 / 025, Loss: 0.983794

* Train accuracy / confusion: 52.38% / [[261, 89, 9], [113, 124, 29], [59, 82, 34]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [11, 20, 4], [6, 16, 1]]

------------------------------ Epoch 035 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.853008
- Iter 024 / 025, Loss: 0.973154

* Train accuracy / confusion: 54.12% / [[259, 83, 16], [113, 121, 34], [55, 66, 53]], 
* Val   accuracy / confusion: 48.08% / [[33, 10, 3], [11, 16, 8], [5, 17, 1]]

------------------------------ Epoch 036 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.858307
- Iter 024 / 025, Loss: 1.048087

* Train accuracy / confusion: 53.38% / [[256, 92, 12], [103, 141, 22], [50, 94, 30]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [7, 20, 8], [4, 18, 1]]

------------------------------ Epoch 037 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.863343
- Iter 024 / 025, Loss: 0.909980

* Train accuracy / confusion: 52.88% / [[259, 82, 15], [113, 114, 43], [56, 68, 50]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 19, 8], [5, 14, 4]]

------------------------------ Epoch 038 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.829956
- Iter 024 / 025, Loss: 0.941467

* Train accuracy / confusion: 53.50% / [[248, 93, 11], [97, 139, 33], [55, 83, 41]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [6, 26, 3], [4, 18, 1]]

------------------------------ Epoch 039 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.018744
- Iter 024 / 025, Loss: 1.040717

* Train accuracy / confusion: 54.38% / [[246, 102, 6], [94, 158, 20], [52, 91, 31]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [11, 16, 8], [6, 13, 4]]

------------------------------ Epoch 040 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.968338
- Iter 024 / 025, Loss: 0.879699

* Train accuracy / confusion: 53.75% / [[260, 86, 11], [107, 129, 33], [59, 74, 41]], 
* Val   accuracy / confusion: 51.92% / [[33, 12, 1], [11, 20, 4], [6, 16, 1]]

------------------------------ Epoch 041 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.997034
- Iter 024 / 025, Loss: 0.785320

* Train accuracy / confusion: 55.00% / [[263, 81, 15], [107, 120, 40], [50, 67, 57]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [8, 19, 8], [4, 18, 1]]

------------------------------ Epoch 042 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.916041
- Iter 024 / 025, Loss: 1.033522

* Train accuracy / confusion: 53.75% / [[249, 99, 11], [105, 140, 26], [49, 80, 41]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [6, 21, 8], [4, 16, 3]]

------------------------------ Epoch 043 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.096227
- Iter 024 / 025, Loss: 0.866128

* Train accuracy / confusion: 53.75% / [[258, 88, 11], [111, 127, 31], [52, 77, 45]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 23, 4], [5, 17, 1]]

------------------------------ Epoch 044 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.826494
- Iter 024 / 025, Loss: 1.098498

* Train accuracy / confusion: 55.38% / [[253, 92, 15], [97, 137, 31], [53, 69, 53]], 
* Val   accuracy / confusion: 53.85% / [[32, 10, 4], [7, 20, 8], [4, 15, 4]]

------------------------------ Epoch 045 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.007441
- Iter 024 / 025, Loss: 0.946291

* Train accuracy / confusion: 54.12% / [[257, 87, 12], [105, 125, 39], [53, 71, 51]], 
* Val   accuracy / confusion: 48.08% / [[33, 10, 3], [12, 16, 7], [7, 15, 1]]

------------------------------ Epoch 046 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.995445
- Iter 024 / 025, Loss: 0.854024

* Train accuracy / confusion: 53.00% / [[241, 111, 4], [99, 156, 12], [51, 99, 27]], 
* Val   accuracy / confusion: 48.08% / [[35, 7, 4], [13, 13, 9], [9, 12, 2]]

------------------------------ Epoch 047 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.001994
- Iter 024 / 025, Loss: 0.958479

* Train accuracy / confusion: 55.62% / [[274, 67, 23], [112, 111, 42], [51, 60, 60]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [7, 22, 6], [4, 18, 1]]

------------------------------ Epoch 048 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.946416
- Iter 024 / 025, Loss: 0.951782

* Train accuracy / confusion: 55.12% / [[249, 103, 7], [89, 161, 19], [42, 99, 31]], 
* Val   accuracy / confusion: 48.08% / [[33, 10, 3], [11, 16, 8], [6, 16, 1]]

------------------------------ Epoch 049 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.881850
- Iter 024 / 025, Loss: 0.942728

* Train accuracy / confusion: 53.12% / [[263, 79, 12], [113, 119, 38], [58, 75, 43]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [6, 21, 8], [4, 17, 2]]

------------------------------ Epoch 050 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.916350
- Iter 024 / 025, Loss: 0.991963

* Train accuracy / confusion: 54.50% / [[250, 93, 13], [102, 138, 29], [49, 78, 48]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 23, 4], [5, 17, 1]]

------------------------------ Epoch 051 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.979981
- Iter 024 / 025, Loss: 0.986771

* Train accuracy / confusion: 52.88% / [[248, 97, 10], [104, 138, 25], [50, 91, 37]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [11, 16, 8], [6, 12, 5]]

------------------------------ Epoch 052 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.851346
- Iter 024 / 025, Loss: 0.915707

* Train accuracy / confusion: 54.50% / [[264, 74, 18], [112, 120, 39], [52, 69, 52]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [7, 20, 8], [4, 18, 1]]

------------------------------ Epoch 053 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.961076
- Iter 024 / 025, Loss: 1.040239

* Train accuracy / confusion: 52.62% / [[248, 101, 9], [111, 125, 30], [46, 82, 48]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [8, 19, 8], [5, 16, 2]]

------------------------------ Epoch 054 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.836649
- Iter 024 / 025, Loss: 0.981696

* Train accuracy / confusion: 54.38% / [[250, 94, 9], [100, 139, 29], [53, 80, 46]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [6, 25, 4], [4, 18, 1]]

------------------------------ Epoch 055 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.080464
- Iter 024 / 025, Loss: 0.811914

* Train accuracy / confusion: 56.00% / [[263, 69, 26], [104, 120, 44], [51, 58, 65]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [7, 20, 8], [4, 16, 3]]

------------------------------ Epoch 056 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.023323
- Iter 024 / 025, Loss: 0.944990

* Train accuracy / confusion: 54.25% / [[253, 96, 8], [100, 143, 27], [51, 84, 38]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [7, 20, 8], [4, 18, 1]]

------------------------------ Epoch 057 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.885690
- Iter 024 / 025, Loss: 0.905349

* Train accuracy / confusion: 53.25% / [[254, 96, 5], [102, 146, 22], [55, 94, 26]], 
* Val   accuracy / confusion: 52.88% / [[33, 8, 5], [8, 17, 10], [5, 13, 5]]

------------------------------ Epoch 058 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.902212
- Iter 024 / 025, Loss: 1.097491

* Train accuracy / confusion: 55.00% / [[257, 90, 12], [98, 128, 40], [50, 70, 55]], 
* Val   accuracy / confusion: 57.69% / [[33, 13, 0], [8, 27, 0], [5, 18, 0]]

------------------------------ Epoch 059 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.941538
- Iter 024 / 025, Loss: 0.920623

* Train accuracy / confusion: 52.75% / [[252, 88, 11], [106, 138, 26], [57, 90, 32]], 
* Val   accuracy / confusion: 50.00% / [[33, 10, 3], [11, 17, 7], [7, 14, 2]]

------------------------------ Epoch 060 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.774454
- Iter 024 / 025, Loss: 1.078500

* Train accuracy / confusion: 54.00% / [[258, 91, 8], [106, 138, 26], [49, 88, 36]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [9, 21, 5], [7, 15, 1]]

------------------------------ Epoch 061 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.894776
- Iter 024 / 025, Loss: 0.925253

* Train accuracy / confusion: 54.88% / [[259, 80, 17], [106, 115, 46], [56, 56, 65]], 
* Val   accuracy / confusion: 53.85% / [[36, 10, 0], [13, 19, 3], [11, 11, 1]]

------------------------------ Epoch 062 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.023841
- Iter 024 / 025, Loss: 1.024187

* Train accuracy / confusion: 54.12% / [[256, 96, 3], [109, 139, 23], [51, 85, 38]], 
* Val   accuracy / confusion: 55.77% / [[34, 11, 1], [8, 22, 5], [5, 16, 2]]

------------------------------ Epoch 063 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.950813
- Iter 024 / 025, Loss: 0.888947

* Train accuracy / confusion: 52.25% / [[253, 95, 9], [104, 132, 30], [52, 92, 33]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [7, 20, 8], [4, 15, 4]]

------------------------------ Epoch 064 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.016331
- Iter 024 / 025, Loss: 0.917247

* Train accuracy / confusion: 52.25% / [[257, 95, 7], [115, 137, 17], [58, 90, 24]], 
* Val   accuracy / confusion: 50.00% / [[35, 9, 2], [13, 16, 6], [9, 13, 1]]

------------------------------ Epoch 065 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.930640
- Iter 024 / 025, Loss: 0.946853

* Train accuracy / confusion: 54.50% / [[272, 70, 16], [120, 112, 38], [56, 64, 52]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 066 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.964787
- Iter 024 / 025, Loss: 1.118938

* Train accuracy / confusion: 55.00% / [[261, 90, 6], [103, 148, 16], [55, 90, 31]], 
* Val   accuracy / confusion: 53.85% / [[32, 9, 5], [6, 19, 10], [4, 14, 5]]

------------------------------ Epoch 067 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.914197
- Iter 024 / 025, Loss: 0.843558

* Train accuracy / confusion: 54.50% / [[237, 96, 21], [94, 136, 45], [45, 63, 63]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [7, 25, 3], [4, 18, 1]]

------------------------------ Epoch 068 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.021743
- Iter 024 / 025, Loss: 0.837760

* Train accuracy / confusion: 55.88% / [[261, 91, 5], [94, 151, 22], [47, 94, 35]], 
* Val   accuracy / confusion: 56.73% / [[32, 14, 0], [7, 26, 2], [4, 18, 1]]

------------------------------ Epoch 069 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.925265
- Iter 024 / 025, Loss: 0.844454

* Train accuracy / confusion: 54.25% / [[255, 88, 14], [103, 132, 31], [56, 74, 47]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [7, 22, 6], [4, 18, 1]]

------------------------------ Epoch 070 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.975890
- Iter 024 / 025, Loss: 1.027460

* Train accuracy / confusion: 54.25% / [[254, 96, 8], [98, 149, 23], [51, 90, 31]], 
* Val   accuracy / confusion: 54.81% / [[32, 13, 1], [7, 24, 4], [4, 18, 1]]

------------------------------ Epoch 071 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.877437
- Iter 024 / 025, Loss: 1.207902

* Train accuracy / confusion: 52.75% / [[248, 87, 19], [105, 122, 42], [54, 71, 52]], 
* Val   accuracy / confusion: 54.81% / [[32, 13, 1], [7, 24, 4], [4, 18, 1]]

------------------------------ Epoch 072 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.970695
- Iter 024 / 025, Loss: 0.893388

* Train accuracy / confusion: 55.38% / [[251, 96, 12], [92, 136, 36], [50, 71, 56]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 23, 4], [5, 17, 1]]

------------------------------ Epoch 073 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.927990
- Iter 024 / 025, Loss: 0.828387

* Train accuracy / confusion: 54.00% / [[242, 104, 10], [98, 148, 24], [49, 83, 42]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [7, 24, 4], [4, 18, 1]]

------------------------------ Epoch 074 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.912010
- Iter 024 / 025, Loss: 0.934519

* Train accuracy / confusion: 55.88% / [[273, 82, 5], [110, 133, 23], [58, 75, 41]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [6, 23, 6], [4, 18, 1]]

------------------------------ Epoch 075 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.832672
- Iter 024 / 025, Loss: 1.000616

* Train accuracy / confusion: 53.12% / [[243, 103, 12], [98, 130, 38], [50, 74, 52]], 
* Val   accuracy / confusion: 55.77% / [[32, 12, 2], [7, 25, 3], [5, 17, 1]]

------------------------------ Epoch 076 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.823258
- Iter 024 / 025, Loss: 0.834370

* Train accuracy / confusion: 54.25% / [[273, 74, 10], [121, 115, 32], [58, 71, 46]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 20, 8], [4, 15, 4]]

------------------------------ Epoch 077 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.997512
- Iter 024 / 025, Loss: 0.786489

* Train accuracy / confusion: 53.75% / [[246, 99, 9], [92, 151, 24], [50, 96, 33]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [7, 22, 6], [4, 18, 1]]

------------------------------ Epoch 078 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.883832
- Iter 024 / 025, Loss: 0.807507

* Train accuracy / confusion: 55.12% / [[255, 88, 13], [105, 136, 30], [50, 73, 50]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [8, 19, 8], [5, 14, 4]]

------------------------------ Epoch 079 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.986806
- Iter 024 / 025, Loss: 0.963553

* Train accuracy / confusion: 54.50% / [[255, 92, 11], [100, 143, 28], [46, 87, 38]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [7, 23, 5], [4, 18, 1]]

------------------------------ Epoch 080 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.917325
- Iter 024 / 025, Loss: 0.916207

* Train accuracy / confusion: 53.62% / [[264, 81, 14], [114, 120, 30], [60, 72, 45]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [8, 21, 6], [4, 18, 1]]

------------------------------ Epoch 081 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.856923
- Iter 024 / 025, Loss: 1.098693

* Train accuracy / confusion: 53.62% / [[255, 97, 7], [98, 135, 31], [54, 84, 39]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [7, 20, 8], [4, 17, 2]]

------------------------------ Epoch 082 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.988418
- Iter 024 / 025, Loss: 0.976813

* Train accuracy / confusion: 53.62% / [[252, 96, 12], [98, 134, 34], [55, 76, 43]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [9, 18, 8], [7, 14, 2]]

------------------------------ Epoch 083 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.920351
- Iter 024 / 025, Loss: 0.847532

* Train accuracy / confusion: 54.75% / [[258, 84, 12], [105, 132, 31], [53, 77, 48]], 
* Val   accuracy / confusion: 50.96% / [[32, 12, 2], [7, 20, 8], [4, 18, 1]]

------------------------------ Epoch 084 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.924249
- Iter 024 / 025, Loss: 0.933679

* Train accuracy / confusion: 52.25% / [[252, 92, 14], [107, 122, 36], [50, 83, 44]], 
* Val   accuracy / confusion: 55.77% / [[32, 13, 1], [7, 25, 3], [4, 18, 1]]

------------------------------ Epoch 085 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.864195
- Iter 024 / 025, Loss: 0.959620

* Train accuracy / confusion: 53.12% / [[248, 103, 5], [103, 136, 32], [50, 82, 41]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [11, 19, 5], [7, 14, 2]]

------------------------------ Epoch 086 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.879821
- Iter 024 / 025, Loss: 0.949800

* Train accuracy / confusion: 54.62% / [[251, 93, 8], [96, 142, 31], [48, 87, 44]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [8, 20, 7], [5, 17, 1]]

------------------------------ Epoch 087 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.861297
- Iter 024 / 025, Loss: 0.950508

* Train accuracy / confusion: 53.25% / [[264, 80, 14], [113, 125, 29], [57, 81, 37]], 
* Val   accuracy / confusion: 54.81% / [[25, 21, 0], [2, 31, 2], [3, 19, 1]]

------------------------------ Epoch 088 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.897090
- Iter 024 / 025, Loss: 0.823656

* Train accuracy / confusion: 53.50% / [[256, 87, 12], [116, 120, 33], [61, 63, 52]], 
* Val   accuracy / confusion: 53.85% / [[32, 9, 5], [6, 19, 10], [4, 14, 5]]

------------------------------ Epoch 089 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.033837
- Iter 024 / 025, Loss: 0.986607

* Train accuracy / confusion: 54.12% / [[246, 95, 12], [90, 147, 31], [46, 93, 40]], 
* Val   accuracy / confusion: 56.73% / [[33, 11, 2], [8, 24, 3], [5, 16, 2]]

------------------------------ Epoch 090 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.899061
- Iter 024 / 025, Loss: 0.921721

* Train accuracy / confusion: 53.88% / [[273, 73, 13], [127, 107, 31], [59, 66, 51]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [7, 18, 10], [4, 14, 5]]

------------------------------ Epoch 091 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.884795
- Iter 024 / 025, Loss: 0.978572

* Train accuracy / confusion: 54.12% / [[245, 98, 11], [95, 135, 38], [48, 77, 53]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [7, 24, 4], [4, 18, 1]]

------------------------------ Epoch 092 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.909856
- Iter 024 / 025, Loss: 0.860619

* Train accuracy / confusion: 54.25% / [[244, 95, 13], [102, 132, 36], [50, 70, 58]], 
* Val   accuracy / confusion: 52.88% / [[28, 17, 1], [5, 26, 4], [4, 18, 1]]

------------------------------ Epoch 093 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.138495
- Iter 024 / 025, Loss: 0.955998

* Train accuracy / confusion: 53.75% / [[235, 113, 6], [90, 158, 23], [50, 88, 37]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [7, 19, 9], [4, 14, 5]]

------------------------------ Epoch 094 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.932562
- Iter 024 / 025, Loss: 0.900473

* Train accuracy / confusion: 53.75% / [[249, 91, 16], [101, 129, 38], [52, 72, 52]], 
* Val   accuracy / confusion: 56.73% / [[31, 15, 0], [6, 27, 2], [4, 18, 1]]

------------------------------ Epoch 095 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.900886
- Iter 024 / 025, Loss: 0.915265

* Train accuracy / confusion: 54.38% / [[240, 102, 9], [85, 160, 26], [46, 97, 35]], 
* Val   accuracy / confusion: 50.96% / [[37, 6, 3], [13, 14, 8], [9, 12, 2]]

------------------------------ Epoch 096 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.046598
- Iter 024 / 025, Loss: 1.024356

* Train accuracy / confusion: 52.25% / [[236, 111, 7], [94, 146, 26], [47, 97, 36]], 
* Val   accuracy / confusion: 52.88% / [[32, 9, 5], [7, 18, 10], [4, 14, 5]]

------------------------------ Epoch 097 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.001694
- Iter 024 / 025, Loss: 0.830829

* Train accuracy / confusion: 54.88% / [[262, 74, 19], [114, 123, 33], [56, 65, 54]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [7, 20, 8], [4, 15, 4]]

------------------------------ Epoch 098 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.077392
- Iter 024 / 025, Loss: 0.983931

* Train accuracy / confusion: 54.12% / [[255, 89, 11], [100, 123, 43], [56, 68, 55]], 
* Val   accuracy / confusion: 54.81% / [[30, 15, 1], [6, 26, 3], [4, 18, 1]]

------------------------------ Epoch 099 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.845833
- Iter 024 / 025, Loss: 1.015669

* Train accuracy / confusion: 52.75% / [[247, 101, 11], [99, 135, 33], [55, 79, 40]], 
* Val   accuracy / confusion: 50.96% / [[35, 8, 3], [13, 16, 6], [8, 13, 2]]

------------------------------ Epoch 100 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.832863
- Iter 024 / 025, Loss: 0.983871

* Train accuracy / confusion: 53.25% / [[241, 98, 14], [102, 140, 27], [53, 80, 45]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [6, 23, 6], [4, 17, 2]]

------------------------------ Epoch 101 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.883571
- Iter 024 / 025, Loss: 0.808283

* Train accuracy / confusion: 55.88% / [[265, 73, 16], [111, 123, 35], [54, 64, 59]], 
* Val   accuracy / confusion: 50.96% / [[37, 6, 3], [13, 13, 9], [8, 12, 3]]

------------------------------ Epoch 102 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.945337
- Iter 024 / 025, Loss: 0.935890

* Train accuracy / confusion: 54.25% / [[267, 86, 5], [112, 139, 18], [53, 92, 28]], 
* Val   accuracy / confusion: 50.00% / [[33, 11, 2], [9, 18, 8], [6, 16, 1]]

------------------------------ Epoch 103 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.987332
- Iter 024 / 025, Loss: 0.919317

* Train accuracy / confusion: 53.38% / [[229, 102, 24], [88, 140, 39], [52, 68, 58]], 
* Val   accuracy / confusion: 51.92% / [[36, 8, 2], [12, 17, 6], [7, 15, 1]]

------------------------------ Epoch 104 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.058399
- Iter 024 / 025, Loss: 1.009620

* Train accuracy / confusion: 54.38% / [[259, 88, 9], [111, 132, 28], [51, 78, 44]], 
* Val   accuracy / confusion: 50.00% / [[28, 16, 2], [5, 23, 7], [4, 18, 1]]

------------------------------ Epoch 105 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.969844
- Iter 024 / 025, Loss: 0.920409

* Train accuracy / confusion: 53.75% / [[254, 93, 7], [115, 134, 24], [56, 75, 42]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [7, 20, 8], [4, 18, 1]]

------------------------------ Epoch 106 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.028869
- Iter 024 / 025, Loss: 0.868347

* Train accuracy / confusion: 54.38% / [[242, 102, 10], [87, 154, 32], [46, 88, 39]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [8, 19, 8], [5, 17, 1]]

------------------------------ Epoch 107 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.907231
- Iter 024 / 025, Loss: 1.012498

* Train accuracy / confusion: 54.00% / [[249, 100, 10], [98, 144, 26], [50, 84, 39]], 
* Val   accuracy / confusion: 51.92% / [[28, 16, 2], [5, 24, 6], [4, 17, 2]]

------------------------------ Epoch 108 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.919066
- Iter 024 / 025, Loss: 0.849067

* Train accuracy / confusion: 55.62% / [[261, 80, 14], [103, 125, 42], [50, 66, 59]], 
* Val   accuracy / confusion: 49.04% / [[33, 11, 2], [11, 17, 7], [6, 16, 1]]

------------------------------ Epoch 109 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.987043
- Iter 024 / 025, Loss: 0.975705

* Train accuracy / confusion: 53.75% / [[240, 102, 15], [97, 136, 35], [46, 75, 54]], 
* Val   accuracy / confusion: 56.73% / [[33, 13, 0], [8, 26, 1], [5, 18, 0]]

------------------------------ Epoch 110 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.968813
- Iter 024 / 025, Loss: 0.990317

* Train accuracy / confusion: 53.50% / [[240, 107, 9], [96, 150, 22], [47, 91, 38]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 19, 8], [6, 14, 3]]

------------------------------ Epoch 111 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.849541
- Iter 024 / 025, Loss: 0.847696

* Train accuracy / confusion: 54.88% / [[279, 68, 11], [120, 108, 38], [58, 66, 52]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [9, 22, 4], [6, 14, 3]]

------------------------------ Epoch 112 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.156584
- Iter 024 / 025, Loss: 0.965232

* Train accuracy / confusion: 55.62% / [[256, 83, 16], [103, 132, 35], [50, 68, 57]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [8, 21, 6], [4, 16, 3]]

------------------------------ Epoch 113 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.903034
- Iter 024 / 025, Loss: 1.010802

* Train accuracy / confusion: 55.25% / [[267, 88, 3], [117, 130, 18], [58, 74, 45]], 
* Val   accuracy / confusion: 55.77% / [[32, 12, 2], [5, 24, 6], [4, 17, 2]]

------------------------------ Epoch 114 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.942684
- Iter 024 / 025, Loss: 1.092889

* Train accuracy / confusion: 53.62% / [[241, 101, 11], [91, 149, 31], [47, 90, 39]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [6, 21, 8], [4, 18, 1]]

------------------------------ Epoch 115 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.960405
- Iter 024 / 025, Loss: 0.952316

* Train accuracy / confusion: 53.62% / [[241, 104, 12], [89, 143, 37], [45, 84, 45]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [8, 24, 3], [5, 17, 1]]

------------------------------ Epoch 116 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.869537
- Iter 024 / 025, Loss: 0.934385

* Train accuracy / confusion: 53.75% / [[276, 78, 5], [113, 136, 16], [53, 105, 18]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 18, 9], [5, 14, 4]]

------------------------------ Epoch 117 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.950862
- Iter 024 / 025, Loss: 0.951250

* Train accuracy / confusion: 53.62% / [[274, 63, 20], [125, 97, 46], [60, 57, 58]], 
* Val   accuracy / confusion: 53.85% / [[37, 7, 2], [13, 18, 4], [8, 14, 1]]

------------------------------ Epoch 118 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.997624
- Iter 024 / 025, Loss: 0.856816

* Train accuracy / confusion: 53.62% / [[256, 94, 8], [111, 128, 29], [56, 73, 45]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [8, 19, 8], [5, 17, 1]]

------------------------------ Epoch 119 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.896823
- Iter 024 / 025, Loss: 0.868899

* Train accuracy / confusion: 53.50% / [[254, 97, 8], [98, 148, 21], [55, 93, 26]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [8, 19, 8], [5, 17, 1]]

------------------------------ Epoch 120 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.885348
- Iter 024 / 025, Loss: 0.940666

* Train accuracy / confusion: 53.50% / [[242, 92, 17], [104, 132, 35], [47, 77, 54]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [10, 18, 7], [5, 16, 2]]

------------------------------ Epoch 121 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.159060
- Iter 024 / 025, Loss: 0.965308

* Train accuracy / confusion: 53.38% / [[250, 95, 13], [106, 132, 28], [50, 81, 45]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [11, 16, 8], [6, 12, 5]]

------------------------------ Epoch 122 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.106238
- Iter 024 / 025, Loss: 0.945849

* Train accuracy / confusion: 54.75% / [[261, 74, 19], [102, 134, 36], [52, 79, 43]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [6, 26, 3], [4, 18, 1]]

------------------------------ Epoch 123 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.933982
- Iter 024 / 025, Loss: 1.068205

* Train accuracy / confusion: 54.50% / [[248, 98, 11], [94, 141, 33], [52, 76, 47]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [11, 16, 8], [6, 12, 5]]

------------------------------ Epoch 124 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.895365
- Iter 024 / 025, Loss: 1.151348

* Train accuracy / confusion: 55.38% / [[262, 72, 22], [112, 108, 46], [55, 50, 73]], 
* Val   accuracy / confusion: 51.92% / [[34, 10, 2], [12, 19, 4], [7, 15, 1]]

------------------------------ Epoch 125 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.843097
- Iter 024 / 025, Loss: 0.987512

* Train accuracy / confusion: 53.38% / [[265, 79, 13], [121, 115, 28], [59, 73, 47]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [7, 20, 8], [4, 18, 1]]

------------------------------ Epoch 126 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.881844
- Iter 024 / 025, Loss: 0.942631

* Train accuracy / confusion: 55.25% / [[236, 105, 12], [82, 161, 28], [45, 86, 45]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [6, 26, 3], [4, 18, 1]]

------------------------------ Epoch 127 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.015773
- Iter 024 / 025, Loss: 0.872390

* Train accuracy / confusion: 55.25% / [[249, 98, 8], [85, 157, 24], [56, 87, 36]], 
* Val   accuracy / confusion: 49.04% / [[28, 15, 3], [6, 21, 8], [4, 17, 2]]

------------------------------ Epoch 128 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.859879
- Iter 024 / 025, Loss: 0.898999

* Train accuracy / confusion: 55.00% / [[254, 89, 15], [95, 135, 35], [52, 74, 51]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [7, 20, 8], [4, 17, 2]]

------------------------------ Epoch 129 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.030303
- Iter 024 / 025, Loss: 0.996281

* Train accuracy / confusion: 54.62% / [[245, 95, 17], [94, 140, 35], [48, 74, 52]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [11, 22, 2], [6, 16, 1]]

------------------------------ Epoch 130 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.917283
- Iter 024 / 025, Loss: 0.907816

* Train accuracy / confusion: 53.75% / [[239, 108, 9], [98, 150, 24], [46, 85, 41]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [6, 23, 6], [4, 17, 2]]

------------------------------ Epoch 131 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.954227
- Iter 024 / 025, Loss: 0.937959

* Train accuracy / confusion: 54.00% / [[270, 76, 13], [125, 109, 32], [59, 63, 53]], 
* Val   accuracy / confusion: 52.88% / [[35, 10, 1], [11, 19, 5], [7, 15, 1]]

------------------------------ Epoch 132 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.780178
- Iter 024 / 025, Loss: 0.846676

* Train accuracy / confusion: 55.12% / [[263, 73, 18], [112, 124, 34], [60, 62, 54]], 
* Val   accuracy / confusion: 53.85% / [[34, 9, 3], [9, 21, 5], [9, 13, 1]]

------------------------------ Epoch 133 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.886647
- Iter 024 / 025, Loss: 0.838110

* Train accuracy / confusion: 54.38% / [[248, 90, 19], [89, 145, 34], [54, 79, 42]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [6, 26, 3], [4, 18, 1]]

------------------------------ Epoch 134 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.967396
- Iter 024 / 025, Loss: 0.903080

* Train accuracy / confusion: 54.12% / [[263, 79, 11], [112, 129, 29], [54, 82, 41]], 
* Val   accuracy / confusion: 57.69% / [[32, 12, 2], [6, 24, 5], [4, 15, 4]]

------------------------------ Epoch 135 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.974201
- Iter 024 / 025, Loss: 0.909306

* Train accuracy / confusion: 53.62% / [[261, 88, 10], [114, 132, 22], [58, 79, 36]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [6, 26, 3], [5, 17, 1]]

------------------------------ Epoch 136 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.922906
- Iter 024 / 025, Loss: 0.869156

* Train accuracy / confusion: 53.50% / [[235, 104, 14], [95, 153, 25], [44, 90, 40]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [9, 17, 9], [5, 12, 6]]

------------------------------ Epoch 137 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.024892
- Iter 024 / 025, Loss: 0.983382

* Train accuracy / confusion: 54.38% / [[263, 80, 11], [108, 127, 35], [55, 76, 45]], 
* Val   accuracy / confusion: 55.77% / [[33, 12, 1], [8, 24, 3], [4, 18, 1]]

------------------------------ Epoch 138 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.940859
- Iter 024 / 025, Loss: 0.918109

* Train accuracy / confusion: 54.50% / [[253, 93, 9], [106, 126, 39], [48, 69, 57]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [7, 21, 7], [4, 18, 1]]

------------------------------ Epoch 139 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.891575
- Iter 024 / 025, Loss: 1.138946

* Train accuracy / confusion: 53.00% / [[274, 71, 11], [135, 107, 27], [63, 69, 43]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [7, 23, 5], [4, 18, 1]]

------------------------------ Epoch 140 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.973055
- Iter 024 / 025, Loss: 0.812702

* Train accuracy / confusion: 53.62% / [[244, 101, 12], [103, 133, 31], [45, 79, 52]], 
* Val   accuracy / confusion: 51.92% / [[35, 8, 3], [13, 14, 8], [9, 9, 5]]

------------------------------ Epoch 141 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.928544
- Iter 024 / 025, Loss: 0.862160

* Train accuracy / confusion: 54.25% / [[261, 81, 14], [107, 122, 36], [58, 70, 51]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 142 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.776459
- Iter 024 / 025, Loss: 0.897546

* Train accuracy / confusion: 53.75% / [[263, 86, 9], [109, 130, 27], [54, 85, 37]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [6, 23, 6], [4, 18, 1]]

------------------------------ Epoch 143 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.112343
- Iter 024 / 025, Loss: 1.048852

* Train accuracy / confusion: 54.75% / [[250, 93, 14], [100, 139, 33], [50, 72, 49]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 144 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.999891
- Iter 024 / 025, Loss: 0.882170

* Train accuracy / confusion: 52.88% / [[231, 106, 18], [86, 151, 33], [46, 88, 41]], 
* Val   accuracy / confusion: 53.85% / [[37, 7, 2], [13, 18, 4], [9, 13, 1]]

------------------------------ Epoch 145 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.933072
- Iter 024 / 025, Loss: 1.064112

* Train accuracy / confusion: 53.25% / [[281, 62, 11], [138, 104, 27], [60, 76, 41]], 
* Val   accuracy / confusion: 50.00% / [[37, 6, 3], [13, 14, 8], [10, 12, 1]]

------------------------------ Epoch 146 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.994292
- Iter 024 / 025, Loss: 0.916700

* Train accuracy / confusion: 53.75% / [[273, 64, 18], [118, 105, 44], [56, 70, 52]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [10, 17, 8], [6, 13, 4]]

------------------------------ Epoch 147 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.983473
- Iter 024 / 025, Loss: 0.872730

* Train accuracy / confusion: 51.25% / [[235, 109, 9], [102, 143, 24], [50, 96, 32]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 23, 5], [4, 18, 1]]

------------------------------ Epoch 148 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.915987
- Iter 024 / 025, Loss: 0.963528

* Train accuracy / confusion: 55.38% / [[268, 67, 23], [109, 111, 43], [56, 59, 64]], 
* Val   accuracy / confusion: 50.96% / [[35, 9, 2], [13, 17, 5], [7, 15, 1]]

------------------------------ Epoch 149 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.948880
- Iter 024 / 025, Loss: 0.966937

* Train accuracy / confusion: 52.38% / [[270, 80, 6], [125, 124, 19], [59, 92, 25]], 
* Val   accuracy / confusion: 48.08% / [[33, 10, 3], [10, 16, 9], [6, 16, 1]]

------------------------------ Epoch 150 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.969228
- Iter 024 / 025, Loss: 0.920507

* Train accuracy / confusion: 53.50% / [[245, 95, 12], [100, 135, 37], [50, 78, 48]], 
* Val   accuracy / confusion: 49.04% / [[34, 11, 1], [13, 16, 6], [8, 14, 1]]

------------------------------ Epoch 151 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.045696
- Iter 024 / 025, Loss: 1.040634

* Train accuracy / confusion: 52.75% / [[246, 96, 12], [104, 132, 32], [51, 83, 44]], 
* Val   accuracy / confusion: 50.00% / [[32, 11, 3], [7, 16, 12], [4, 15, 4]]

------------------------------ Epoch 152 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.926184
- Iter 024 / 025, Loss: 0.894245

* Train accuracy / confusion: 52.25% / [[258, 84, 11], [115, 118, 38], [54, 80, 42]], 
* Val   accuracy / confusion: 49.04% / [[33, 10, 3], [11, 16, 8], [6, 15, 2]]

------------------------------ Epoch 153 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.978096
- Iter 024 / 025, Loss: 0.969169

* Train accuracy / confusion: 53.62% / [[279, 56, 24], [136, 83, 48], [63, 44, 67]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [7, 22, 6], [4, 18, 1]]

------------------------------ Epoch 154 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.985022
- Iter 024 / 025, Loss: 1.026893

* Train accuracy / confusion: 55.88% / [[258, 91, 6], [102, 141, 27], [53, 74, 48]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [11, 20, 4], [6, 16, 1]]

------------------------------ Epoch 155 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.886322
- Iter 024 / 025, Loss: 1.165070

* Train accuracy / confusion: 54.62% / [[263, 80, 10], [114, 130, 28], [53, 78, 44]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [7, 21, 7], [5, 16, 2]]

------------------------------ Epoch 156 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.993022
- Iter 024 / 025, Loss: 0.887316

* Train accuracy / confusion: 55.12% / [[249, 88, 18], [103, 139, 29], [48, 73, 53]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [7, 22, 6], [5, 15, 3]]

------------------------------ Epoch 157 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.898682
- Iter 024 / 025, Loss: 0.890749

* Train accuracy / confusion: 53.88% / [[247, 93, 19], [99, 140, 30], [43, 85, 44]], 
* Val   accuracy / confusion: 50.96% / [[29, 14, 3], [5, 20, 10], [4, 15, 4]]

------------------------------ Epoch 158 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.731315
- Iter 024 / 025, Loss: 0.931985

* Train accuracy / confusion: 53.62% / [[246, 95, 11], [100, 136, 35], [53, 77, 47]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 23, 4], [4, 18, 1]]

------------------------------ Epoch 159 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.075369
- Iter 024 / 025, Loss: 0.856424

* Train accuracy / confusion: 54.88% / [[244, 100, 11], [94, 143, 31], [47, 78, 52]], 
* Val   accuracy / confusion: 50.00% / [[34, 10, 2], [11, 17, 7], [7, 15, 1]]

------------------------------ Epoch 160 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.842879
- Iter 024 / 025, Loss: 1.004213

* Train accuracy / confusion: 54.62% / [[268, 77, 14], [114, 129, 25], [54, 79, 40]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [7, 21, 7], [5, 14, 4]]

------------------------------ Epoch 161 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.035168
- Iter 024 / 025, Loss: 0.949668

* Train accuracy / confusion: 54.50% / [[244, 99, 14], [99, 124, 42], [56, 54, 68]], 
* Val   accuracy / confusion: 50.96% / [[35, 9, 2], [13, 17, 5], [8, 14, 1]]

------------------------------ Epoch 162 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.025392
- Iter 024 / 025, Loss: 0.965270

* Train accuracy / confusion: 54.00% / [[258, 86, 11], [111, 131, 27], [52, 81, 43]], 
* Val   accuracy / confusion: 50.00% / [[33, 9, 4], [10, 17, 8], [6, 15, 2]]

------------------------------ Epoch 163 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.874323
- Iter 024 / 025, Loss: 0.946707

* Train accuracy / confusion: 53.50% / [[252, 91, 9], [103, 132, 38], [56, 75, 44]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [6, 22, 7], [4, 18, 1]]

------------------------------ Epoch 164 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.940419
- Iter 024 / 025, Loss: 0.835167

* Train accuracy / confusion: 55.88% / [[246, 98, 11], [85, 151, 32], [44, 83, 50]], 
* Val   accuracy / confusion: 50.96% / [[33, 9, 4], [10, 18, 7], [5, 16, 2]]

------------------------------ Epoch 165 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.008664
- Iter 024 / 025, Loss: 0.956042

* Train accuracy / confusion: 54.62% / [[268, 80, 11], [108, 123, 36], [63, 65, 46]], 
* Val   accuracy / confusion: 51.92% / [[33, 12, 1], [11, 20, 4], [6, 16, 1]]

------------------------------ Epoch 166 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.096785
- Iter 024 / 025, Loss: 1.094092

* Train accuracy / confusion: 52.88% / [[278, 69, 12], [127, 100, 37], [55, 77, 45]], 
* Val   accuracy / confusion: 50.96% / [[33, 11, 2], [11, 19, 5], [6, 16, 1]]

------------------------------ Epoch 167 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.949977
- Iter 024 / 025, Loss: 1.067932

* Train accuracy / confusion: 54.12% / [[265, 81, 9], [113, 124, 32], [55, 77, 44]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [11, 18, 6], [5, 15, 3]]

------------------------------ Epoch 168 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.843556
- Iter 024 / 025, Loss: 1.083348

* Train accuracy / confusion: 52.62% / [[263, 89, 3], [117, 120, 29], [57, 84, 38]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 169 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.830304
- Iter 024 / 025, Loss: 0.834940

* Train accuracy / confusion: 55.25% / [[258, 80, 17], [108, 125, 38], [55, 60, 59]], 
* Val   accuracy / confusion: 55.77% / [[32, 10, 4], [7, 22, 6], [4, 15, 4]]

------------------------------ Epoch 170 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.969497
- Iter 024 / 025, Loss: 0.990201

* Train accuracy / confusion: 54.88% / [[270, 74, 14], [113, 119, 36], [57, 67, 50]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [8, 19, 8], [5, 14, 4]]

------------------------------ Epoch 171 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.036474
- Iter 024 / 025, Loss: 0.975813

* Train accuracy / confusion: 53.62% / [[252, 85, 16], [105, 128, 38], [51, 76, 49]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [7, 25, 3], [4, 18, 1]]

------------------------------ Epoch 172 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.932449
- Iter 024 / 025, Loss: 1.049118

* Train accuracy / confusion: 52.50% / [[249, 95, 15], [105, 125, 35], [54, 76, 46]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [10, 18, 7], [6, 13, 4]]

------------------------------ Epoch 173 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.986096
- Iter 024 / 025, Loss: 0.917933

* Train accuracy / confusion: 54.38% / [[267, 75, 14], [116, 120, 31], [59, 70, 48]], 
* Val   accuracy / confusion: 50.00% / [[34, 9, 3], [12, 15, 8], [6, 14, 3]]

------------------------------ Epoch 174 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.933542
- Iter 024 / 025, Loss: 1.164733

* Train accuracy / confusion: 54.00% / [[259, 88, 8], [111, 127, 32], [53, 76, 46]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [8, 20, 7], [5, 14, 4]]

------------------------------ Epoch 175 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.051715
- Iter 024 / 025, Loss: 1.010500

* Train accuracy / confusion: 54.38% / [[267, 77, 10], [110, 121, 35], [55, 78, 47]], 
* Val   accuracy / confusion: 48.08% / [[35, 7, 4], [13, 13, 9], [9, 12, 2]]

------------------------------ Epoch 176 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.040079
- Iter 024 / 025, Loss: 1.064026

* Train accuracy / confusion: 52.50% / [[275, 72, 8], [133, 106, 30], [58, 79, 39]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 23, 4], [5, 17, 1]]

------------------------------ Epoch 177 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.095198
- Iter 024 / 025, Loss: 1.003438

* Train accuracy / confusion: 53.62% / [[280, 59, 18], [132, 92, 43], [63, 56, 57]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [7, 23, 5], [4, 17, 2]]

------------------------------ Epoch 178 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.888648
- Iter 024 / 025, Loss: 0.945857

* Train accuracy / confusion: 54.38% / [[244, 96, 6], [102, 146, 25], [52, 84, 45]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [6, 20, 9], [4, 17, 2]]

------------------------------ Epoch 179 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.026169
- Iter 024 / 025, Loss: 0.871921

* Train accuracy / confusion: 54.50% / [[262, 71, 22], [112, 113, 41], [56, 62, 61]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 180 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.858293
- Iter 024 / 025, Loss: 0.828222

* Train accuracy / confusion: 55.88% / [[259, 85, 14], [97, 132, 39], [52, 66, 56]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [8, 18, 9], [5, 15, 3]]

------------------------------ Epoch 181 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.999920
- Iter 024 / 025, Loss: 0.991739

* Train accuracy / confusion: 55.25% / [[252, 95, 8], [95, 157, 17], [54, 89, 33]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [6, 26, 3], [4, 18, 1]]

------------------------------ Epoch 182 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.787510
- Iter 024 / 025, Loss: 0.926109

* Train accuracy / confusion: 54.75% / [[252, 93, 10], [104, 131, 33], [54, 68, 55]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 22, 5], [5, 16, 2]]

------------------------------ Epoch 183 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.119786
- Iter 024 / 025, Loss: 1.125282

* Train accuracy / confusion: 54.88% / [[246, 95, 13], [99, 146, 27], [46, 81, 47]], 
* Val   accuracy / confusion: 53.85% / [[25, 19, 2], [2, 30, 3], [3, 19, 1]]

------------------------------ Epoch 184 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.945349
- Iter 024 / 025, Loss: 1.006542

* Train accuracy / confusion: 54.25% / [[242, 101, 10], [95, 149, 25], [44, 91, 43]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [11, 15, 9], [6, 12, 5]]

------------------------------ Epoch 185 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.955012
- Iter 024 / 025, Loss: 1.024997

* Train accuracy / confusion: 54.62% / [[272, 71, 12], [112, 118, 40], [56, 72, 47]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [7, 23, 5], [4, 18, 1]]

------------------------------ Epoch 186 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.967037
- Iter 024 / 025, Loss: 0.886307

* Train accuracy / confusion: 53.88% / [[259, 91, 8], [101, 132, 35], [48, 86, 40]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [11, 21, 3], [5, 17, 1]]

------------------------------ Epoch 187 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.086895
- Iter 024 / 025, Loss: 0.811423

* Train accuracy / confusion: 55.12% / [[261, 84, 14], [110, 125, 32], [52, 67, 55]], 
* Val   accuracy / confusion: 50.00% / [[33, 11, 2], [11, 17, 7], [6, 15, 2]]

------------------------------ Epoch 188 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.989550
- Iter 024 / 025, Loss: 0.956069

* Train accuracy / confusion: 55.25% / [[259, 83, 16], [103, 131, 39], [54, 63, 52]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [7, 26, 2], [4, 18, 1]]

------------------------------ Epoch 189 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.126519
- Iter 024 / 025, Loss: 0.840631

* Train accuracy / confusion: 53.38% / [[262, 83, 12], [114, 114, 38], [56, 70, 51]], 
* Val   accuracy / confusion: 51.92% / [[35, 8, 3], [13, 17, 5], [8, 13, 2]]

------------------------------ Epoch 190 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.839543
- Iter 024 / 025, Loss: 0.896605

* Train accuracy / confusion: 53.38% / [[268, 80, 8], [116, 126, 30], [51, 88, 33]], 
* Val   accuracy / confusion: 50.96% / [[28, 16, 2], [5, 23, 7], [4, 17, 2]]

------------------------------ Epoch 191 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.898341
- Iter 024 / 025, Loss: 0.813120

* Train accuracy / confusion: 54.12% / [[248, 96, 11], [103, 140, 27], [52, 78, 45]], 
* Val   accuracy / confusion: 52.88% / [[34, 8, 4], [12, 17, 6], [6, 13, 4]]

------------------------------ Epoch 192 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.959150
- Iter 024 / 025, Loss: 0.994287

* Train accuracy / confusion: 55.12% / [[272, 69, 16], [119, 116, 31], [60, 64, 53]], 
* Val   accuracy / confusion: 53.85% / [[28, 16, 2], [5, 26, 4], [4, 17, 2]]

------------------------------ Epoch 193 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.970507
- Iter 024 / 025, Loss: 0.981744

* Train accuracy / confusion: 54.00% / [[253, 88, 15], [105, 120, 41], [56, 63, 59]], 
* Val   accuracy / confusion: 57.69% / [[32, 13, 1], [6, 27, 2], [4, 18, 1]]

------------------------------ Epoch 194 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.893964
- Iter 024 / 025, Loss: 0.932807

* Train accuracy / confusion: 51.88% / [[251, 99, 9], [107, 140, 20], [52, 98, 24]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [10, 21, 4], [6, 14, 3]]

------------------------------ Epoch 195 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.076905
- Iter 024 / 025, Loss: 0.828321

* Train accuracy / confusion: 57.12% / [[263, 78, 17], [100, 136, 33], [46, 69, 58]], 
* Val   accuracy / confusion: 58.65% / [[36, 7, 3], [7, 20, 8], [5, 13, 5]]

------------------------------ Epoch 196 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.001861
- Iter 024 / 025, Loss: 1.052124

* Train accuracy / confusion: 53.88% / [[257, 84, 16], [104, 133, 30], [51, 84, 41]], 
* Val   accuracy / confusion: 54.81% / [[31, 12, 3], [7, 23, 5], [4, 16, 3]]

------------------------------ Epoch 197 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.999630
- Iter 024 / 025, Loss: 1.096680

* Train accuracy / confusion: 55.00% / [[257, 75, 22], [112, 116, 42], [55, 54, 67]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [8, 17, 10], [5, 14, 4]]

------------------------------ Epoch 198 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.933759
- Iter 024 / 025, Loss: 0.995723

* Train accuracy / confusion: 54.62% / [[259, 92, 8], [106, 140, 22], [53, 82, 38]], 
* Val   accuracy / confusion: 50.00% / [[33, 11, 2], [10, 18, 7], [5, 17, 1]]

------------------------------ Epoch 199 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 1.029096
- Iter 024 / 025, Loss: 1.032506

* Train accuracy / confusion: 55.25% / [[265, 78, 16], [111, 122, 37], [53, 63, 55]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [6, 23, 6], [4, 18, 1]]

------------------------------ Epoch 200 / 500, Learning rate: 3.28e-02 ------------------------------
- Iter 012 / 025, Loss: 0.872324
- Iter 024 / 025, Loss: 0.935604

* Train accuracy / confusion: 54.62% / [[239, 99, 12], [99, 138, 34], [45, 74, 60]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 201 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.895055
- Iter 024 / 025, Loss: 0.963016

* Train accuracy / confusion: 56.00% / [[257, 88, 14], [101, 139, 29], [47, 73, 52]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [8, 23, 4], [5, 16, 2]]

------------------------------ Epoch 202 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.887994
- Iter 024 / 025, Loss: 1.136903

* Train accuracy / confusion: 55.00% / [[258, 92, 7], [100, 136, 32], [46, 83, 46]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [7, 23, 5], [4, 17, 2]]

------------------------------ Epoch 203 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.851268
- Iter 024 / 025, Loss: 0.863900

* Train accuracy / confusion: 54.00% / [[251, 90, 14], [102, 132, 31], [50, 81, 49]], 
* Val   accuracy / confusion: 55.77% / [[32, 10, 4], [7, 24, 4], [4, 17, 2]]

------------------------------ Epoch 204 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.044032
- Iter 024 / 025, Loss: 1.029680

* Train accuracy / confusion: 55.88% / [[251, 92, 14], [93, 143, 32], [49, 73, 53]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 205 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.917099
- Iter 024 / 025, Loss: 0.869908

* Train accuracy / confusion: 55.12% / [[261, 87, 11], [99, 136, 33], [52, 77, 44]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 206 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.825882
- Iter 024 / 025, Loss: 1.012437

* Train accuracy / confusion: 55.12% / [[258, 88, 9], [108, 135, 26], [52, 76, 48]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 207 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.772555
- Iter 024 / 025, Loss: 0.796726

* Train accuracy / confusion: 55.00% / [[253, 88, 13], [107, 141, 26], [49, 77, 46]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [7, 23, 5], [5, 16, 2]]

------------------------------ Epoch 208 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.066664
- Iter 024 / 025, Loss: 1.079190

* Train accuracy / confusion: 54.62% / [[249, 93, 10], [98, 136, 36], [48, 78, 52]], 
* Val   accuracy / confusion: 50.96% / [[32, 10, 4], [7, 20, 8], [4, 18, 1]]

------------------------------ Epoch 209 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.823209
- Iter 024 / 025, Loss: 1.217807

* Train accuracy / confusion: 56.00% / [[258, 91, 8], [93, 144, 30], [46, 84, 46]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [7, 23, 5], [5, 16, 2]]

------------------------------ Epoch 210 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.963458
- Iter 024 / 025, Loss: 0.902949

* Train accuracy / confusion: 56.00% / [[250, 98, 10], [91, 145, 32], [48, 73, 53]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [7, 21, 7], [4, 18, 1]]

------------------------------ Epoch 211 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.758969
- Iter 024 / 025, Loss: 0.896388

* Train accuracy / confusion: 54.25% / [[253, 89, 10], [105, 136, 30], [48, 84, 45]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 212 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.124456
- Iter 024 / 025, Loss: 1.002283

* Train accuracy / confusion: 54.62% / [[256, 92, 11], [103, 130, 31], [52, 74, 51]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [9, 21, 5], [5, 16, 2]]

------------------------------ Epoch 213 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.069165
- Iter 024 / 025, Loss: 0.875896

* Train accuracy / confusion: 55.38% / [[261, 84, 12], [108, 130, 24], [55, 74, 52]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 214 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.846761
- Iter 024 / 025, Loss: 0.901846

* Train accuracy / confusion: 55.75% / [[260, 90, 10], [108, 131, 28], [51, 67, 55]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 215 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.938981
- Iter 024 / 025, Loss: 1.091016

* Train accuracy / confusion: 55.88% / [[256, 90, 12], [105, 137, 26], [49, 71, 54]], 
* Val   accuracy / confusion: 50.00% / [[33, 11, 2], [9, 18, 8], [5, 17, 1]]

------------------------------ Epoch 216 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.000644
- Iter 024 / 025, Loss: 0.884760

* Train accuracy / confusion: 55.88% / [[261, 83, 10], [112, 126, 32], [50, 66, 60]], 
* Val   accuracy / confusion: 50.96% / [[33, 10, 3], [8, 19, 8], [5, 17, 1]]

------------------------------ Epoch 217 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.044379
- Iter 024 / 025, Loss: 0.969604

* Train accuracy / confusion: 55.25% / [[263, 79, 20], [114, 120, 32], [47, 66, 59]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [9, 20, 6], [5, 16, 2]]

------------------------------ Epoch 218 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.987606
- Iter 024 / 025, Loss: 1.004360

* Train accuracy / confusion: 54.75% / [[257, 84, 15], [106, 127, 37], [52, 68, 54]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [10, 19, 6], [5, 16, 2]]

------------------------------ Epoch 219 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.895429
- Iter 024 / 025, Loss: 0.944596

* Train accuracy / confusion: 55.50% / [[248, 86, 18], [105, 140, 24], [52, 71, 56]], 
* Val   accuracy / confusion: 56.73% / [[33, 10, 3], [8, 23, 4], [5, 15, 3]]

------------------------------ Epoch 220 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.951994
- Iter 024 / 025, Loss: 0.967427

* Train accuracy / confusion: 54.88% / [[252, 90, 11], [104, 133, 34], [54, 68, 54]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [8, 21, 6], [5, 15, 3]]

------------------------------ Epoch 221 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.843957
- Iter 024 / 025, Loss: 1.244883

* Train accuracy / confusion: 55.88% / [[253, 95, 11], [99, 136, 30], [52, 66, 58]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 222 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.989967
- Iter 024 / 025, Loss: 1.002919

* Train accuracy / confusion: 56.00% / [[260, 88, 8], [97, 141, 31], [49, 79, 47]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [8, 20, 7], [5, 17, 1]]

------------------------------ Epoch 223 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.932232
- Iter 024 / 025, Loss: 0.912418

* Train accuracy / confusion: 56.25% / [[261, 83, 13], [103, 135, 32], [53, 66, 54]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 224 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.771885
- Iter 024 / 025, Loss: 0.841282

* Train accuracy / confusion: 56.12% / [[259, 87, 10], [104, 137, 26], [57, 67, 53]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 20, 7], [5, 15, 3]]

------------------------------ Epoch 225 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.022227
- Iter 024 / 025, Loss: 0.906368

* Train accuracy / confusion: 55.12% / [[265, 87, 12], [104, 127, 27], [53, 76, 49]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 226 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.881536
- Iter 024 / 025, Loss: 0.873643

* Train accuracy / confusion: 54.38% / [[254, 95, 9], [102, 132, 36], [47, 76, 49]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [7, 21, 7], [5, 15, 3]]

------------------------------ Epoch 227 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.875895
- Iter 024 / 025, Loss: 0.900486

* Train accuracy / confusion: 55.62% / [[258, 81, 14], [108, 135, 26], [59, 67, 52]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 20, 7], [5, 14, 4]]

------------------------------ Epoch 228 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.958632
- Iter 024 / 025, Loss: 0.824783

* Train accuracy / confusion: 53.50% / [[257, 88, 11], [111, 125, 31], [55, 76, 46]], 
* Val   accuracy / confusion: 50.96% / [[33, 12, 1], [9, 19, 7], [6, 16, 1]]

------------------------------ Epoch 229 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.907726
- Iter 024 / 025, Loss: 0.959968

* Train accuracy / confusion: 55.88% / [[259, 88, 12], [102, 136, 30], [47, 74, 52]], 
* Val   accuracy / confusion: 50.96% / [[32, 12, 2], [8, 19, 8], [5, 16, 2]]

------------------------------ Epoch 230 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.067995
- Iter 024 / 025, Loss: 0.881927

* Train accuracy / confusion: 55.50% / [[257, 90, 11], [103, 137, 28], [52, 72, 50]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 231 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.968782
- Iter 024 / 025, Loss: 1.099724

* Train accuracy / confusion: 55.12% / [[255, 91, 9], [105, 135, 30], [49, 75, 51]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [8, 22, 5], [5, 16, 2]]

------------------------------ Epoch 232 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.873196
- Iter 024 / 025, Loss: 0.779258

* Train accuracy / confusion: 56.38% / [[254, 97, 9], [95, 143, 24], [51, 73, 54]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [8, 21, 6], [5, 15, 3]]

------------------------------ Epoch 233 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.100102
- Iter 024 / 025, Loss: 0.938508

* Train accuracy / confusion: 54.88% / [[255, 84, 16], [103, 127, 36], [50, 72, 57]], 
* Val   accuracy / confusion: 52.88% / [[32, 9, 5], [7, 21, 7], [5, 16, 2]]

------------------------------ Epoch 234 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.981522
- Iter 024 / 025, Loss: 1.067313

* Train accuracy / confusion: 54.88% / [[257, 84, 15], [112, 124, 33], [52, 65, 58]], 
* Val   accuracy / confusion: 50.00% / [[33, 11, 2], [9, 17, 9], [5, 16, 2]]

------------------------------ Epoch 235 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.888231
- Iter 024 / 025, Loss: 0.899028

* Train accuracy / confusion: 56.25% / [[256, 87, 14], [110, 125, 33], [52, 54, 69]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [9, 22, 4], [5, 17, 1]]

------------------------------ Epoch 236 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.917475
- Iter 024 / 025, Loss: 0.896712

* Train accuracy / confusion: 53.75% / [[251, 90, 14], [99, 128, 43], [53, 71, 51]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [6, 22, 7], [4, 16, 3]]

------------------------------ Epoch 237 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.914415
- Iter 024 / 025, Loss: 0.912542

* Train accuracy / confusion: 56.88% / [[256, 94, 6], [95, 146, 30], [48, 72, 53]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [7, 22, 6], [5, 16, 2]]

------------------------------ Epoch 238 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.660307
- Iter 024 / 025, Loss: 0.963909

* Train accuracy / confusion: 54.62% / [[250, 92, 13], [100, 139, 30], [51, 77, 48]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 239 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.776703
- Iter 024 / 025, Loss: 0.943204

* Train accuracy / confusion: 57.25% / [[254, 95, 8], [96, 144, 25], [52, 66, 60]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 240 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.808028
- Iter 024 / 025, Loss: 0.831057

* Train accuracy / confusion: 55.38% / [[257, 89, 12], [102, 130, 36], [51, 67, 56]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 241 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.992404
- Iter 024 / 025, Loss: 1.089292

* Train accuracy / confusion: 55.88% / [[260, 80, 15], [105, 130, 33], [50, 70, 57]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 242 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.947622
- Iter 024 / 025, Loss: 0.907284

* Train accuracy / confusion: 56.00% / [[250, 92, 10], [102, 143, 28], [52, 68, 55]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 19, 8], [5, 15, 3]]

------------------------------ Epoch 243 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.947139
- Iter 024 / 025, Loss: 0.994584

* Train accuracy / confusion: 54.75% / [[250, 94, 13], [103, 133, 29], [55, 68, 55]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [8, 18, 9], [5, 14, 4]]

------------------------------ Epoch 244 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.024396
- Iter 024 / 025, Loss: 0.839897

* Train accuracy / confusion: 56.62% / [[256, 87, 12], [99, 146, 26], [53, 70, 51]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [8, 20, 7], [5, 15, 3]]

------------------------------ Epoch 245 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.853353
- Iter 024 / 025, Loss: 0.899720

* Train accuracy / confusion: 55.75% / [[256, 86, 15], [104, 135, 30], [50, 69, 55]], 
* Val   accuracy / confusion: 50.96% / [[33, 11, 2], [8, 18, 9], [5, 16, 2]]

------------------------------ Epoch 246 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.966993
- Iter 024 / 025, Loss: 1.043338

* Train accuracy / confusion: 56.25% / [[263, 85, 10], [104, 131, 34], [49, 68, 56]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [8, 20, 7], [6, 14, 3]]

------------------------------ Epoch 247 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.795026
- Iter 024 / 025, Loss: 1.042653

* Train accuracy / confusion: 52.38% / [[257, 95, 7], [110, 111, 41], [58, 70, 51]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [9, 22, 4], [5, 15, 3]]

------------------------------ Epoch 248 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.855559
- Iter 024 / 025, Loss: 0.832082

* Train accuracy / confusion: 55.12% / [[255, 89, 12], [97, 137, 32], [49, 80, 49]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [7, 19, 9], [5, 13, 5]]

------------------------------ Epoch 249 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.920036
- Iter 024 / 025, Loss: 0.909942

* Train accuracy / confusion: 57.50% / [[260, 85, 11], [99, 139, 29], [52, 64, 61]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 250 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.912596
- Iter 024 / 025, Loss: 0.924619

* Train accuracy / confusion: 55.38% / [[258, 91, 6], [107, 130, 32], [47, 74, 55]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 251 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.983032
- Iter 024 / 025, Loss: 0.885620

* Train accuracy / confusion: 56.75% / [[257, 92, 10], [93, 138, 35], [51, 65, 59]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 252 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.980047
- Iter 024 / 025, Loss: 1.019717

* Train accuracy / confusion: 56.62% / [[254, 87, 12], [100, 139, 30], [47, 71, 60]], 
* Val   accuracy / confusion: 50.96% / [[33, 11, 2], [8, 18, 9], [5, 16, 2]]

------------------------------ Epoch 253 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.899040
- Iter 024 / 025, Loss: 1.011362

* Train accuracy / confusion: 54.88% / [[252, 92, 13], [100, 135, 32], [54, 70, 52]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 254 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.962984
- Iter 024 / 025, Loss: 0.945283

* Train accuracy / confusion: 56.12% / [[247, 97, 11], [95, 143, 29], [54, 65, 59]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 255 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.916474
- Iter 024 / 025, Loss: 1.071566

* Train accuracy / confusion: 55.38% / [[241, 97, 11], [104, 144, 23], [52, 70, 58]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [8, 22, 5], [5, 16, 2]]

------------------------------ Epoch 256 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.040064
- Iter 024 / 025, Loss: 0.818381

* Train accuracy / confusion: 53.00% / [[244, 98, 13], [102, 130, 40], [52, 71, 50]], 
* Val   accuracy / confusion: 56.73% / [[33, 10, 3], [8, 21, 6], [5, 13, 5]]

------------------------------ Epoch 257 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.895194
- Iter 024 / 025, Loss: 0.813377

* Train accuracy / confusion: 53.62% / [[240, 103, 9], [99, 138, 34], [50, 76, 51]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [6, 21, 8], [4, 16, 3]]

------------------------------ Epoch 258 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.158449
- Iter 024 / 025, Loss: 0.874818

* Train accuracy / confusion: 57.25% / [[245, 97, 11], [91, 153, 25], [55, 63, 60]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [6, 23, 6], [4, 18, 1]]

------------------------------ Epoch 259 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.037262
- Iter 024 / 025, Loss: 0.954488

* Train accuracy / confusion: 54.38% / [[243, 102, 9], [100, 137, 34], [47, 73, 55]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 260 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.855182
- Iter 024 / 025, Loss: 0.884681

* Train accuracy / confusion: 56.00% / [[246, 97, 14], [98, 141, 27], [49, 67, 61]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 20, 7], [5, 15, 3]]

------------------------------ Epoch 261 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.083736
- Iter 024 / 025, Loss: 1.029792

* Train accuracy / confusion: 57.12% / [[253, 90, 13], [91, 145, 33], [46, 70, 59]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 262 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.986256
- Iter 024 / 025, Loss: 1.363945

* Train accuracy / confusion: 53.88% / [[259, 88, 12], [107, 122, 35], [53, 74, 50]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [9, 19, 7], [5, 15, 3]]

------------------------------ Epoch 263 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.028222
- Iter 024 / 025, Loss: 0.816284

* Train accuracy / confusion: 55.38% / [[252, 91, 10], [107, 132, 31], [54, 64, 59]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [8, 21, 6], [5, 15, 3]]

------------------------------ Epoch 264 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.905884
- Iter 024 / 025, Loss: 0.781464

* Train accuracy / confusion: 56.38% / [[259, 89, 10], [101, 138, 28], [56, 65, 54]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [9, 22, 4], [6, 15, 2]]

------------------------------ Epoch 265 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.021206
- Iter 024 / 025, Loss: 0.910950

* Train accuracy / confusion: 55.00% / [[253, 90, 15], [101, 132, 32], [51, 71, 55]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 21, 6], [5, 17, 1]]

------------------------------ Epoch 266 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.829961
- Iter 024 / 025, Loss: 0.931928

* Train accuracy / confusion: 59.00% / [[266, 83, 8], [95, 145, 29], [54, 59, 61]], 
* Val   accuracy / confusion: 49.04% / [[33, 10, 3], [11, 17, 7], [5, 17, 1]]

------------------------------ Epoch 267 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.911712
- Iter 024 / 025, Loss: 0.965916

* Train accuracy / confusion: 54.38% / [[259, 89, 10], [106, 123, 39], [52, 69, 53]], 
* Val   accuracy / confusion: 51.92% / [[33, 9, 4], [10, 19, 6], [5, 16, 2]]

------------------------------ Epoch 268 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.008546
- Iter 024 / 025, Loss: 0.785208

* Train accuracy / confusion: 56.50% / [[256, 91, 13], [105, 138, 26], [54, 59, 58]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [8, 19, 8], [5, 15, 3]]

------------------------------ Epoch 269 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.027907
- Iter 024 / 025, Loss: 0.811255

* Train accuracy / confusion: 57.00% / [[263, 83, 9], [104, 132, 29], [53, 66, 61]], 
* Val   accuracy / confusion: 56.73% / [[33, 11, 2], [8, 22, 5], [5, 14, 4]]

------------------------------ Epoch 270 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.890130
- Iter 024 / 025, Loss: 1.124723

* Train accuracy / confusion: 54.62% / [[252, 91, 12], [111, 132, 29], [54, 66, 53]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 271 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.843929
- Iter 024 / 025, Loss: 0.999958

* Train accuracy / confusion: 57.00% / [[254, 91, 8], [100, 141, 31], [50, 64, 61]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 272 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.867406
- Iter 024 / 025, Loss: 0.738986

* Train accuracy / confusion: 55.50% / [[250, 92, 13], [98, 141, 31], [50, 72, 53]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [8, 21, 6], [5, 14, 4]]

------------------------------ Epoch 273 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.834396
- Iter 024 / 025, Loss: 0.988605

* Train accuracy / confusion: 56.75% / [[261, 84, 11], [102, 132, 32], [51, 66, 61]], 
* Val   accuracy / confusion: 51.92% / [[33, 11, 2], [8, 19, 8], [5, 16, 2]]

------------------------------ Epoch 274 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.104239
- Iter 024 / 025, Loss: 0.923561

* Train accuracy / confusion: 55.50% / [[253, 94, 11], [99, 133, 33], [49, 70, 58]], 
* Val   accuracy / confusion: 57.69% / [[32, 13, 1], [6, 25, 4], [4, 16, 3]]

------------------------------ Epoch 275 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.969585
- Iter 024 / 025, Loss: 0.945665

* Train accuracy / confusion: 57.12% / [[262, 92, 7], [95, 135, 41], [48, 60, 60]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [7, 20, 8], [5, 15, 3]]

------------------------------ Epoch 276 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.997778
- Iter 024 / 025, Loss: 0.912868

* Train accuracy / confusion: 54.75% / [[240, 106, 6], [99, 139, 31], [50, 70, 59]], 
* Val   accuracy / confusion: 51.92% / [[33, 8, 5], [8, 17, 10], [5, 14, 4]]

------------------------------ Epoch 277 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.888587
- Iter 024 / 025, Loss: 1.018275

* Train accuracy / confusion: 57.25% / [[249, 95, 10], [96, 147, 28], [48, 65, 62]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 278 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.209211
- Iter 024 / 025, Loss: 0.749742

* Train accuracy / confusion: 57.00% / [[255, 90, 10], [100, 139, 30], [48, 66, 62]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [8, 22, 5], [4, 17, 2]]

------------------------------ Epoch 279 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.876776
- Iter 024 / 025, Loss: 0.913124

* Train accuracy / confusion: 54.38% / [[252, 92, 11], [103, 128, 34], [54, 71, 55]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 280 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.917743
- Iter 024 / 025, Loss: 0.862926

* Train accuracy / confusion: 55.38% / [[254, 91, 8], [107, 131, 31], [54, 66, 58]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 281 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.937543
- Iter 024 / 025, Loss: 0.827498

* Train accuracy / confusion: 55.25% / [[253, 91, 12], [102, 139, 32], [47, 74, 50]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [8, 20, 7], [5, 14, 4]]

------------------------------ Epoch 282 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.910462
- Iter 024 / 025, Loss: 0.842102

* Train accuracy / confusion: 55.88% / [[255, 89, 11], [100, 129, 38], [56, 59, 63]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 283 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.886146
- Iter 024 / 025, Loss: 0.982260

* Train accuracy / confusion: 55.62% / [[256, 90, 12], [102, 132, 34], [52, 65, 57]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 284 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.890688
- Iter 024 / 025, Loss: 0.996576

* Train accuracy / confusion: 56.25% / [[251, 98, 6], [101, 136, 30], [53, 62, 63]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 285 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.938720
- Iter 024 / 025, Loss: 0.991666

* Train accuracy / confusion: 54.88% / [[261, 86, 13], [105, 124, 35], [55, 67, 54]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [9, 22, 4], [5, 16, 2]]

------------------------------ Epoch 286 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.956269
- Iter 024 / 025, Loss: 1.082256

* Train accuracy / confusion: 57.38% / [[258, 86, 13], [101, 137, 30], [53, 58, 64]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 287 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.906039
- Iter 024 / 025, Loss: 1.097486

* Train accuracy / confusion: 54.38% / [[250, 100, 8], [103, 123, 39], [50, 65, 62]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [6, 23, 6], [4, 15, 4]]

------------------------------ Epoch 288 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.064145
- Iter 024 / 025, Loss: 0.915315

* Train accuracy / confusion: 56.75% / [[252, 95, 7], [96, 136, 33], [51, 64, 66]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [8, 19, 8], [5, 14, 4]]

------------------------------ Epoch 289 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.923736
- Iter 024 / 025, Loss: 1.024296

* Train accuracy / confusion: 55.25% / [[248, 94, 12], [100, 138, 35], [51, 66, 56]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [6, 23, 6], [4, 17, 2]]

------------------------------ Epoch 290 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.883923
- Iter 024 / 025, Loss: 0.919156

* Train accuracy / confusion: 54.88% / [[242, 93, 16], [93, 147, 33], [50, 76, 50]], 
* Val   accuracy / confusion: 55.77% / [[32, 10, 4], [6, 23, 6], [4, 16, 3]]

------------------------------ Epoch 291 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.997909
- Iter 024 / 025, Loss: 0.871034

* Train accuracy / confusion: 57.12% / [[246, 97, 10], [93, 148, 31], [52, 60, 63]], 
* Val   accuracy / confusion: 54.81% / [[32, 12, 2], [7, 21, 7], [4, 15, 4]]

------------------------------ Epoch 292 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.949530
- Iter 024 / 025, Loss: 0.862415

* Train accuracy / confusion: 57.25% / [[238, 109, 7], [91, 155, 24], [49, 62, 65]], 
* Val   accuracy / confusion: 52.88% / [[31, 11, 4], [6, 22, 7], [4, 17, 2]]

------------------------------ Epoch 293 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.971169
- Iter 024 / 025, Loss: 0.981656

* Train accuracy / confusion: 55.38% / [[256, 98, 6], [97, 134, 36], [49, 71, 53]], 
* Val   accuracy / confusion: 50.96% / [[33, 12, 1], [8, 18, 9], [5, 16, 2]]

------------------------------ Epoch 294 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.815534
- Iter 024 / 025, Loss: 0.983894

* Train accuracy / confusion: 57.12% / [[261, 91, 9], [98, 137, 30], [57, 58, 59]], 
* Val   accuracy / confusion: 56.73% / [[34, 9, 3], [8, 22, 5], [6, 14, 3]]

------------------------------ Epoch 295 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.937605
- Iter 024 / 025, Loss: 0.962712

* Train accuracy / confusion: 55.88% / [[252, 90, 12], [100, 134, 34], [52, 65, 61]], 
* Val   accuracy / confusion: 55.77% / [[34, 10, 2], [8, 21, 6], [4, 16, 3]]

------------------------------ Epoch 296 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.972873
- Iter 024 / 025, Loss: 0.910126

* Train accuracy / confusion: 56.75% / [[252, 92, 12], [96, 142, 33], [54, 59, 60]], 
* Val   accuracy / confusion: 56.73% / [[33, 10, 3], [6, 23, 6], [5, 15, 3]]

------------------------------ Epoch 297 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.909336
- Iter 024 / 025, Loss: 0.982282

* Train accuracy / confusion: 57.00% / [[253, 93, 9], [99, 145, 27], [48, 68, 58]], 
* Val   accuracy / confusion: 53.85% / [[33, 11, 2], [7, 21, 7], [5, 16, 2]]

------------------------------ Epoch 298 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.906627
- Iter 024 / 025, Loss: 0.827651

* Train accuracy / confusion: 56.25% / [[258, 88, 8], [98, 136, 33], [51, 72, 56]], 
* Val   accuracy / confusion: 55.77% / [[33, 12, 1], [7, 23, 5], [4, 17, 2]]

------------------------------ Epoch 299 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.791661
- Iter 024 / 025, Loss: 0.850612

* Train accuracy / confusion: 57.88% / [[255, 93, 10], [96, 141, 29], [54, 55, 67]], 
* Val   accuracy / confusion: 50.96% / [[33, 11, 2], [8, 18, 9], [5, 16, 2]]

------------------------------ Epoch 300 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.945165
- Iter 024 / 025, Loss: 0.956147

* Train accuracy / confusion: 55.38% / [[251, 93, 10], [111, 128, 30], [52, 61, 64]], 
* Val   accuracy / confusion: 53.85% / [[33, 8, 5], [8, 20, 7], [5, 15, 3]]

------------------------------ Epoch 301 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.924224
- Iter 024 / 025, Loss: 1.010102

* Train accuracy / confusion: 55.75% / [[247, 92, 15], [98, 141, 30], [45, 74, 58]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [6, 20, 9], [4, 16, 3]]

------------------------------ Epoch 302 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.017139
- Iter 024 / 025, Loss: 0.838080

* Train accuracy / confusion: 56.25% / [[256, 92, 10], [100, 131, 35], [52, 61, 63]], 
* Val   accuracy / confusion: 56.73% / [[33, 8, 5], [8, 22, 5], [5, 14, 4]]

------------------------------ Epoch 303 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.816932
- Iter 024 / 025, Loss: 1.002169

* Train accuracy / confusion: 56.12% / [[256, 80, 20], [106, 134, 28], [49, 68, 59]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [8, 21, 6], [5, 15, 3]]

------------------------------ Epoch 304 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.855303
- Iter 024 / 025, Loss: 0.874116

* Train accuracy / confusion: 56.62% / [[238, 104, 11], [94, 149, 27], [48, 63, 66]], 
* Val   accuracy / confusion: 55.77% / [[32, 9, 5], [6, 24, 5], [4, 17, 2]]

------------------------------ Epoch 305 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.055637
- Iter 024 / 025, Loss: 0.794436

* Train accuracy / confusion: 57.00% / [[240, 100, 13], [96, 147, 30], [53, 52, 69]], 
* Val   accuracy / confusion: 56.73% / [[33, 10, 3], [8, 24, 3], [5, 16, 2]]

------------------------------ Epoch 306 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.846485
- Iter 024 / 025, Loss: 0.702590

* Train accuracy / confusion: 56.88% / [[244, 104, 11], [94, 148, 25], [51, 60, 63]], 
* Val   accuracy / confusion: 51.92% / [[32, 9, 5], [6, 20, 9], [4, 17, 2]]

------------------------------ Epoch 307 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.952071
- Iter 024 / 025, Loss: 0.850884

* Train accuracy / confusion: 56.62% / [[261, 85, 12], [104, 128, 37], [52, 57, 64]], 
* Val   accuracy / confusion: 50.96% / [[33, 9, 4], [8, 18, 9], [5, 16, 2]]

------------------------------ Epoch 308 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.853378
- Iter 024 / 025, Loss: 0.942367

* Train accuracy / confusion: 55.00% / [[245, 97, 12], [104, 134, 34], [51, 62, 61]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [6, 22, 7], [4, 17, 2]]

------------------------------ Epoch 309 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.989182
- Iter 024 / 025, Loss: 1.018219

* Train accuracy / confusion: 54.62% / [[253, 89, 13], [109, 130, 29], [54, 69, 54]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [7, 22, 6], [5, 16, 2]]

------------------------------ Epoch 310 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.830436
- Iter 024 / 025, Loss: 1.048167

* Train accuracy / confusion: 55.88% / [[257, 88, 11], [103, 130, 34], [51, 66, 60]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 19, 8], [5, 15, 3]]

------------------------------ Epoch 311 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.871758
- Iter 024 / 025, Loss: 0.961148

* Train accuracy / confusion: 56.62% / [[256, 83, 13], [104, 135, 32], [50, 65, 62]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [8, 22, 5], [5, 17, 1]]

------------------------------ Epoch 312 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.013918
- Iter 024 / 025, Loss: 0.865990

* Train accuracy / confusion: 56.25% / [[250, 96, 9], [97, 142, 33], [52, 63, 58]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [7, 21, 7], [5, 14, 4]]

------------------------------ Epoch 313 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.008234
- Iter 024 / 025, Loss: 0.909127

* Train accuracy / confusion: 55.75% / [[249, 101, 8], [95, 140, 32], [45, 73, 57]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [8, 21, 6], [5, 16, 2]]

------------------------------ Epoch 314 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.766289
- Iter 024 / 025, Loss: 0.828839

* Train accuracy / confusion: 57.12% / [[244, 102, 11], [95, 144, 31], [51, 53, 69]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [6, 23, 6], [4, 18, 1]]

------------------------------ Epoch 315 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.795620
- Iter 024 / 025, Loss: 1.101367

* Train accuracy / confusion: 54.75% / [[247, 103, 8], [101, 134, 31], [54, 65, 57]], 
* Val   accuracy / confusion: 51.92% / [[32, 12, 2], [9, 19, 7], [5, 15, 3]]

------------------------------ Epoch 316 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.014735
- Iter 024 / 025, Loss: 0.723372

* Train accuracy / confusion: 58.00% / [[260, 86, 15], [93, 145, 30], [54, 58, 59]], 
* Val   accuracy / confusion: 54.81% / [[33, 8, 5], [8, 21, 6], [5, 15, 3]]

------------------------------ Epoch 317 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.866164
- Iter 024 / 025, Loss: 0.936776

* Train accuracy / confusion: 55.25% / [[261, 88, 12], [112, 125, 29], [58, 59, 56]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [8, 23, 4], [6, 15, 2]]

------------------------------ Epoch 318 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.891836
- Iter 024 / 025, Loss: 0.902132

* Train accuracy / confusion: 58.25% / [[260, 88, 9], [100, 138, 30], [52, 55, 68]], 
* Val   accuracy / confusion: 52.88% / [[33, 10, 3], [9, 20, 6], [5, 16, 2]]

------------------------------ Epoch 319 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.875159
- Iter 024 / 025, Loss: 1.042795

* Train accuracy / confusion: 56.00% / [[259, 85, 13], [105, 132, 33], [50, 66, 57]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 320 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.913639
- Iter 024 / 025, Loss: 1.123451

* Train accuracy / confusion: 54.75% / [[252, 98, 8], [109, 129, 32], [53, 62, 57]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [10, 22, 3], [5, 16, 2]]

------------------------------ Epoch 321 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.957884
- Iter 024 / 025, Loss: 0.912496

* Train accuracy / confusion: 55.62% / [[245, 99, 12], [99, 140, 29], [52, 64, 60]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [7, 23, 5], [4, 17, 2]]

------------------------------ Epoch 322 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.762460
- Iter 024 / 025, Loss: 0.986399

* Train accuracy / confusion: 55.00% / [[248, 94, 13], [97, 137, 34], [50, 72, 55]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 20, 7], [5, 16, 2]]

------------------------------ Epoch 323 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.883115
- Iter 024 / 025, Loss: 0.915517

* Train accuracy / confusion: 55.25% / [[242, 100, 12], [96, 145, 28], [54, 68, 55]], 
* Val   accuracy / confusion: 56.73% / [[33, 11, 2], [7, 20, 8], [5, 12, 6]]

------------------------------ Epoch 324 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.040856
- Iter 024 / 025, Loss: 0.993900

* Train accuracy / confusion: 55.50% / [[256, 92, 8], [101, 131, 34], [50, 71, 57]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [8, 19, 8], [5, 15, 3]]

------------------------------ Epoch 325 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.959599
- Iter 024 / 025, Loss: 0.860646

* Train accuracy / confusion: 55.88% / [[254, 94, 13], [103, 135, 28], [50, 65, 58]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [8, 21, 6], [5, 15, 3]]

------------------------------ Epoch 326 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.136621
- Iter 024 / 025, Loss: 1.121392

* Train accuracy / confusion: 55.88% / [[262, 84, 15], [102, 131, 35], [49, 68, 54]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [9, 22, 4], [4, 16, 3]]

------------------------------ Epoch 327 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.947635
- Iter 024 / 025, Loss: 0.831851

* Train accuracy / confusion: 57.12% / [[255, 92, 8], [99, 138, 30], [54, 60, 64]], 
* Val   accuracy / confusion: 53.85% / [[32, 10, 4], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 328 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.896586
- Iter 024 / 025, Loss: 0.967641

* Train accuracy / confusion: 55.62% / [[253, 95, 7], [103, 133, 31], [54, 65, 59]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [8, 22, 5], [5, 15, 3]]

------------------------------ Epoch 329 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.986784
- Iter 024 / 025, Loss: 0.859154

* Train accuracy / confusion: 58.38% / [[253, 93, 12], [98, 149, 25], [48, 57, 65]], 
* Val   accuracy / confusion: 55.77% / [[32, 9, 5], [6, 23, 6], [4, 16, 3]]

------------------------------ Epoch 330 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.767470
- Iter 024 / 025, Loss: 0.842452

* Train accuracy / confusion: 56.88% / [[245, 96, 12], [97, 148, 27], [50, 63, 62]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 331 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.805889
- Iter 024 / 025, Loss: 0.759897

* Train accuracy / confusion: 56.12% / [[253, 89, 14], [98, 137, 32], [51, 67, 59]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [8, 21, 6], [4, 16, 3]]

------------------------------ Epoch 332 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.018942
- Iter 024 / 025, Loss: 0.896762

* Train accuracy / confusion: 58.25% / [[261, 80, 18], [93, 138, 31], [46, 66, 67]], 
* Val   accuracy / confusion: 58.65% / [[33, 11, 2], [7, 23, 5], [5, 13, 5]]

------------------------------ Epoch 333 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.886831
- Iter 024 / 025, Loss: 0.888345

* Train accuracy / confusion: 56.50% / [[258, 88, 13], [98, 131, 38], [49, 62, 63]], 
* Val   accuracy / confusion: 53.85% / [[33, 8, 5], [7, 19, 9], [4, 15, 4]]

------------------------------ Epoch 334 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.765640
- Iter 024 / 025, Loss: 1.075397

* Train accuracy / confusion: 57.00% / [[256, 93, 12], [96, 140, 32], [52, 59, 60]], 
* Val   accuracy / confusion: 53.85% / [[33, 9, 4], [9, 20, 6], [4, 16, 3]]

------------------------------ Epoch 335 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.928823
- Iter 024 / 025, Loss: 0.798998

* Train accuracy / confusion: 57.50% / [[260, 87, 14], [105, 130, 28], [51, 55, 70]], 
* Val   accuracy / confusion: 53.85% / [[34, 11, 1], [9, 19, 7], [4, 16, 3]]

------------------------------ Epoch 336 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.856795
- Iter 024 / 025, Loss: 0.971203

* Train accuracy / confusion: 55.12% / [[252, 92, 12], [106, 130, 32], [52, 65, 59]], 
* Val   accuracy / confusion: 55.77% / [[33, 10, 3], [9, 22, 4], [4, 16, 3]]

------------------------------ Epoch 337 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.850564
- Iter 024 / 025, Loss: 0.794141

* Train accuracy / confusion: 56.75% / [[249, 97, 12], [94, 144, 30], [53, 60, 61]], 
* Val   accuracy / confusion: 54.81% / [[33, 12, 1], [7, 21, 7], [4, 16, 3]]

------------------------------ Epoch 338 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.965277
- Iter 024 / 025, Loss: 0.835191

* Train accuracy / confusion: 56.12% / [[249, 95, 15], [98, 141, 28], [51, 64, 59]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [8, 22, 5], [5, 15, 3]]

------------------------------ Epoch 339 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.884669
- Iter 024 / 025, Loss: 1.148452

* Train accuracy / confusion: 55.25% / [[246, 96, 14], [101, 130, 37], [50, 60, 66]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [8, 20, 7], [4, 17, 2]]

------------------------------ Epoch 340 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.989852
- Iter 024 / 025, Loss: 0.853921

* Train accuracy / confusion: 55.25% / [[249, 99, 7], [101, 132, 35], [50, 66, 61]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [7, 23, 5], [4, 15, 4]]

------------------------------ Epoch 341 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.923555
- Iter 024 / 025, Loss: 0.846484

* Train accuracy / confusion: 58.75% / [[254, 86, 14], [92, 149, 30], [50, 58, 67]], 
* Val   accuracy / confusion: 52.88% / [[32, 14, 0], [7, 21, 7], [5, 16, 2]]

------------------------------ Epoch 342 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.825909
- Iter 024 / 025, Loss: 0.908617

* Train accuracy / confusion: 57.62% / [[257, 92, 11], [92, 146, 28], [52, 64, 58]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [8, 20, 7], [4, 18, 1]]

------------------------------ Epoch 343 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.929062
- Iter 024 / 025, Loss: 0.848957

* Train accuracy / confusion: 56.25% / [[260, 87, 11], [99, 131, 32], [55, 66, 59]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [9, 20, 6], [5, 16, 2]]

------------------------------ Epoch 344 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.895905
- Iter 024 / 025, Loss: 0.822600

* Train accuracy / confusion: 57.62% / [[253, 82, 17], [99, 139, 33], [50, 58, 69]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [7, 22, 6], [4, 15, 4]]

------------------------------ Epoch 345 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.950259
- Iter 024 / 025, Loss: 0.808932

* Train accuracy / confusion: 57.50% / [[257, 86, 17], [97, 135, 30], [52, 58, 68]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [8, 20, 7], [5, 17, 1]]

------------------------------ Epoch 346 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.783906
- Iter 024 / 025, Loss: 0.889939

* Train accuracy / confusion: 57.12% / [[252, 90, 13], [100, 140, 30], [49, 61, 65]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [6, 20, 9], [5, 16, 2]]

------------------------------ Epoch 347 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.859493
- Iter 024 / 025, Loss: 0.859052

* Train accuracy / confusion: 57.25% / [[244, 98, 12], [101, 143, 27], [51, 53, 71]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [7, 21, 7], [4, 16, 3]]

------------------------------ Epoch 348 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.881440
- Iter 024 / 025, Loss: 0.826968

* Train accuracy / confusion: 56.25% / [[248, 103, 12], [97, 136, 31], [53, 54, 66]], 
* Val   accuracy / confusion: 53.85% / [[32, 10, 4], [7, 23, 5], [4, 18, 1]]

------------------------------ Epoch 349 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.063326
- Iter 024 / 025, Loss: 0.883890

* Train accuracy / confusion: 56.62% / [[248, 91, 12], [99, 147, 27], [52, 66, 58]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [8, 21, 6], [5, 14, 4]]

------------------------------ Epoch 350 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.975106
- Iter 024 / 025, Loss: 0.870460

* Train accuracy / confusion: 56.00% / [[244, 95, 14], [97, 143, 28], [50, 68, 61]], 
* Val   accuracy / confusion: 52.88% / [[33, 11, 2], [8, 20, 7], [4, 17, 2]]

------------------------------ Epoch 351 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.849460
- Iter 024 / 025, Loss: 1.056023

* Train accuracy / confusion: 56.88% / [[248, 98, 12], [100, 141, 28], [50, 57, 66]], 
* Val   accuracy / confusion: 55.77% / [[32, 9, 5], [7, 23, 5], [5, 15, 3]]

------------------------------ Epoch 352 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.990581
- Iter 024 / 025, Loss: 0.672982

* Train accuracy / confusion: 55.75% / [[254, 87, 15], [99, 133, 36], [53, 64, 59]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [7, 20, 8], [4, 17, 2]]

------------------------------ Epoch 353 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.933823
- Iter 024 / 025, Loss: 0.915013

* Train accuracy / confusion: 56.62% / [[245, 97, 10], [97, 145, 29], [50, 64, 63]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [7, 24, 4], [4, 18, 1]]

------------------------------ Epoch 354 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.046253
- Iter 024 / 025, Loss: 0.944448

* Train accuracy / confusion: 57.88% / [[253, 87, 13], [96, 148, 25], [52, 64, 62]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [7, 21, 7], [5, 15, 3]]

------------------------------ Epoch 355 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.047263
- Iter 024 / 025, Loss: 0.922112

* Train accuracy / confusion: 57.25% / [[250, 93, 11], [99, 139, 32], [54, 53, 69]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [7, 21, 7], [4, 16, 3]]

------------------------------ Epoch 356 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.834631
- Iter 024 / 025, Loss: 0.899220

* Train accuracy / confusion: 56.12% / [[252, 97, 9], [102, 133, 32], [49, 62, 64]], 
* Val   accuracy / confusion: 50.96% / [[33, 8, 5], [8, 17, 10], [4, 16, 3]]

------------------------------ Epoch 357 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.920977
- Iter 024 / 025, Loss: 0.855296

* Train accuracy / confusion: 56.12% / [[251, 84, 20], [92, 141, 33], [52, 70, 57]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [7, 21, 7], [4, 17, 2]]

------------------------------ Epoch 358 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.966810
- Iter 024 / 025, Loss: 0.797798

* Train accuracy / confusion: 56.25% / [[255, 87, 17], [106, 131, 29], [51, 60, 64]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [7, 21, 7], [4, 15, 4]]

------------------------------ Epoch 359 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.798871
- Iter 024 / 025, Loss: 0.809768

* Train accuracy / confusion: 56.88% / [[248, 92, 14], [109, 133, 27], [48, 55, 74]], 
* Val   accuracy / confusion: 57.69% / [[33, 10, 3], [7, 24, 4], [5, 15, 3]]

------------------------------ Epoch 360 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.821027
- Iter 024 / 025, Loss: 0.996887

* Train accuracy / confusion: 57.38% / [[248, 98, 7], [97, 139, 34], [54, 51, 72]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [7, 21, 7], [4, 15, 4]]

------------------------------ Epoch 361 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.874466
- Iter 024 / 025, Loss: 0.799311

* Train accuracy / confusion: 58.00% / [[252, 90, 14], [92, 144, 32], [45, 63, 68]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [6, 22, 7], [4, 17, 2]]

------------------------------ Epoch 362 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.053164
- Iter 024 / 025, Loss: 0.984408

* Train accuracy / confusion: 58.00% / [[252, 90, 15], [92, 148, 29], [48, 62, 64]], 
* Val   accuracy / confusion: 54.81% / [[31, 12, 3], [6, 23, 6], [4, 16, 3]]

------------------------------ Epoch 363 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.825663
- Iter 024 / 025, Loss: 0.933166

* Train accuracy / confusion: 58.25% / [[248, 97, 12], [91, 149, 26], [47, 61, 69]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [6, 23, 6], [4, 16, 3]]

------------------------------ Epoch 364 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.075830
- Iter 024 / 025, Loss: 1.042897

* Train accuracy / confusion: 55.12% / [[238, 106, 11], [92, 143, 34], [48, 68, 60]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 20, 8], [4, 15, 4]]

------------------------------ Epoch 365 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.891396
- Iter 024 / 025, Loss: 1.001372

* Train accuracy / confusion: 55.75% / [[244, 96, 13], [96, 139, 33], [50, 66, 63]], 
* Val   accuracy / confusion: 52.88% / [[30, 12, 4], [7, 22, 6], [4, 16, 3]]

------------------------------ Epoch 366 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.256512
- Iter 024 / 025, Loss: 0.973696

* Train accuracy / confusion: 56.12% / [[242, 95, 14], [100, 139, 31], [48, 63, 68]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 367 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.856716
- Iter 024 / 025, Loss: 1.020346

* Train accuracy / confusion: 56.62% / [[245, 99, 11], [94, 144, 27], [48, 68, 64]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [6, 21, 8], [4, 17, 2]]

------------------------------ Epoch 368 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.925432
- Iter 024 / 025, Loss: 1.128406

* Train accuracy / confusion: 57.00% / [[248, 91, 13], [88, 147, 33], [53, 66, 61]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 20, 8], [4, 15, 4]]

------------------------------ Epoch 369 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.904112
- Iter 024 / 025, Loss: 0.957630

* Train accuracy / confusion: 56.88% / [[248, 101, 12], [90, 142, 31], [54, 57, 65]], 
* Val   accuracy / confusion: 57.69% / [[33, 12, 1], [7, 25, 3], [4, 17, 2]]

------------------------------ Epoch 370 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.908643
- Iter 024 / 025, Loss: 0.773568

* Train accuracy / confusion: 58.12% / [[263, 88, 8], [102, 136, 23], [54, 60, 66]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [7, 18, 10], [4, 16, 3]]

------------------------------ Epoch 371 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.022122
- Iter 024 / 025, Loss: 0.951778

* Train accuracy / confusion: 59.38% / [[262, 74, 19], [101, 139, 30], [52, 49, 74]], 
* Val   accuracy / confusion: 55.77% / [[33, 12, 1], [9, 21, 5], [4, 15, 4]]

------------------------------ Epoch 372 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.857615
- Iter 024 / 025, Loss: 0.794807

* Train accuracy / confusion: 57.12% / [[248, 97, 10], [98, 140, 34], [48, 56, 69]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [6, 21, 8], [4, 17, 2]]

------------------------------ Epoch 373 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.819055
- Iter 024 / 025, Loss: 0.887376

* Train accuracy / confusion: 59.00% / [[255, 91, 12], [90, 149, 29], [52, 54, 68]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [6, 23, 6], [6, 14, 3]]

------------------------------ Epoch 374 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.012496
- Iter 024 / 025, Loss: 0.912996

* Train accuracy / confusion: 56.62% / [[261, 83, 9], [103, 131, 34], [52, 66, 61]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 375 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.873348
- Iter 024 / 025, Loss: 0.786432

* Train accuracy / confusion: 57.75% / [[254, 87, 13], [91, 142, 37], [52, 58, 66]], 
* Val   accuracy / confusion: 55.77% / [[31, 12, 3], [6, 24, 5], [4, 16, 3]]

------------------------------ Epoch 376 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.080074
- Iter 024 / 025, Loss: 0.953393

* Train accuracy / confusion: 57.88% / [[246, 98, 11], [91, 147, 32], [51, 54, 70]], 
* Val   accuracy / confusion: 51.92% / [[31, 13, 2], [7, 21, 7], [5, 16, 2]]

------------------------------ Epoch 377 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.188303
- Iter 024 / 025, Loss: 0.744186

* Train accuracy / confusion: 56.62% / [[246, 100, 8], [98, 137, 34], [50, 57, 70]], 
* Val   accuracy / confusion: 54.81% / [[32, 9, 5], [8, 23, 4], [4, 17, 2]]

------------------------------ Epoch 378 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.828689
- Iter 024 / 025, Loss: 0.849997

* Train accuracy / confusion: 58.25% / [[258, 91, 10], [99, 137, 29], [49, 56, 71]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [6, 20, 9], [4, 16, 3]]

------------------------------ Epoch 379 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.884076
- Iter 024 / 025, Loss: 0.880881

* Train accuracy / confusion: 56.88% / [[262, 78, 12], [105, 127, 38], [50, 62, 66]], 
* Val   accuracy / confusion: 56.73% / [[32, 13, 1], [6, 24, 5], [5, 15, 3]]

------------------------------ Epoch 380 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.858081
- Iter 024 / 025, Loss: 0.989169

* Train accuracy / confusion: 58.88% / [[270, 81, 8], [106, 136, 27], [50, 57, 65]], 
* Val   accuracy / confusion: 54.81% / [[33, 9, 4], [8, 21, 6], [6, 14, 3]]

------------------------------ Epoch 381 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.910135
- Iter 024 / 025, Loss: 1.130472

* Train accuracy / confusion: 56.75% / [[249, 87, 18], [104, 142, 25], [52, 60, 63]], 
* Val   accuracy / confusion: 49.04% / [[30, 11, 5], [7, 18, 10], [4, 16, 3]]

------------------------------ Epoch 382 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.884502
- Iter 024 / 025, Loss: 0.820568

* Train accuracy / confusion: 57.50% / [[258, 87, 10], [101, 137, 32], [54, 56, 65]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 383 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.981399
- Iter 024 / 025, Loss: 0.845524

* Train accuracy / confusion: 57.62% / [[258, 87, 11], [103, 136, 28], [51, 59, 67]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [7, 21, 7], [5, 14, 4]]

------------------------------ Epoch 384 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.861834
- Iter 024 / 025, Loss: 0.959842

* Train accuracy / confusion: 58.50% / [[263, 84, 15], [94, 138, 34], [51, 54, 67]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [8, 21, 6], [6, 13, 4]]

------------------------------ Epoch 385 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.782307
- Iter 024 / 025, Loss: 1.037890

* Train accuracy / confusion: 57.62% / [[252, 90, 14], [95, 142, 30], [51, 59, 67]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [7, 21, 7], [4, 16, 3]]

------------------------------ Epoch 386 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.938760
- Iter 024 / 025, Loss: 0.820400

* Train accuracy / confusion: 58.38% / [[255, 86, 12], [98, 144, 27], [44, 66, 68]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [7, 20, 8], [4, 17, 2]]

------------------------------ Epoch 387 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.913465
- Iter 024 / 025, Loss: 1.102370

* Train accuracy / confusion: 58.38% / [[258, 89, 10], [95, 140, 30], [55, 54, 69]], 
* Val   accuracy / confusion: 55.77% / [[31, 11, 4], [7, 22, 6], [4, 14, 5]]

------------------------------ Epoch 388 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.943705
- Iter 024 / 025, Loss: 0.922424

* Train accuracy / confusion: 59.00% / [[256, 87, 12], [89, 146, 31], [53, 56, 70]], 
* Val   accuracy / confusion: 54.81% / [[32, 10, 4], [6, 23, 6], [5, 16, 2]]

------------------------------ Epoch 389 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.900707
- Iter 024 / 025, Loss: 0.885890

* Train accuracy / confusion: 57.62% / [[257, 87, 8], [100, 135, 35], [50, 59, 69]], 
* Val   accuracy / confusion: 53.85% / [[33, 12, 1], [7, 20, 8], [6, 14, 3]]

------------------------------ Epoch 390 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.917411
- Iter 024 / 025, Loss: 0.867925

* Train accuracy / confusion: 56.38% / [[253, 89, 15], [99, 130, 37], [50, 59, 68]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [6, 20, 9], [5, 14, 4]]

------------------------------ Epoch 391 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.750115
- Iter 024 / 025, Loss: 1.043094

* Train accuracy / confusion: 56.88% / [[247, 95, 14], [91, 141, 35], [53, 57, 67]], 
* Val   accuracy / confusion: 54.81% / [[31, 14, 1], [6, 23, 6], [5, 15, 3]]

------------------------------ Epoch 392 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.027591
- Iter 024 / 025, Loss: 0.856341

* Train accuracy / confusion: 58.75% / [[254, 89, 15], [86, 152, 28], [47, 65, 64]], 
* Val   accuracy / confusion: 56.73% / [[33, 12, 1], [7, 23, 5], [6, 14, 3]]

------------------------------ Epoch 393 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 1.083483
- Iter 024 / 025, Loss: 1.019772

* Train accuracy / confusion: 58.62% / [[256, 92, 11], [86, 150, 32], [49, 61, 63]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [7, 21, 7], [4, 17, 2]]

------------------------------ Epoch 394 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.789063
- Iter 024 / 025, Loss: 0.982304

* Train accuracy / confusion: 58.25% / [[262, 85, 13], [90, 145, 31], [49, 66, 59]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 395 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.854379
- Iter 024 / 025, Loss: 0.999155

* Train accuracy / confusion: 58.50% / [[259, 84, 14], [100, 147, 23], [49, 62, 62]], 
* Val   accuracy / confusion: 54.81% / [[30, 14, 2], [6, 24, 5], [4, 16, 3]]

------------------------------ Epoch 396 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.784232
- Iter 024 / 025, Loss: 0.992000

* Train accuracy / confusion: 57.25% / [[245, 104, 8], [93, 139, 30], [50, 57, 74]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [6, 22, 7], [4, 16, 3]]

------------------------------ Epoch 397 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.897066
- Iter 024 / 025, Loss: 0.957846

* Train accuracy / confusion: 59.00% / [[256, 83, 17], [96, 149, 27], [47, 58, 67]], 
* Val   accuracy / confusion: 53.85% / [[31, 10, 5], [6, 23, 6], [4, 17, 2]]

------------------------------ Epoch 398 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.756127
- Iter 024 / 025, Loss: 0.919720

* Train accuracy / confusion: 56.75% / [[249, 100, 11], [97, 133, 34], [46, 58, 72]], 
* Val   accuracy / confusion: 57.69% / [[33, 12, 1], [6, 24, 5], [4, 16, 3]]

------------------------------ Epoch 399 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.984358
- Iter 024 / 025, Loss: 0.879830

* Train accuracy / confusion: 59.00% / [[262, 75, 20], [96, 139, 33], [47, 57, 71]], 
* Val   accuracy / confusion: 54.81% / [[33, 11, 2], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 400 / 500, Learning rate: 3.28e-03 ------------------------------
- Iter 012 / 025, Loss: 0.837836
- Iter 024 / 025, Loss: 1.032981

* Train accuracy / confusion: 57.88% / [[251, 95, 10], [94, 143, 29], [53, 56, 69]], 
* Val   accuracy / confusion: 52.88% / [[30, 13, 3], [6, 21, 8], [4, 15, 4]]

------------------------------ Epoch 401 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.942831
- Iter 024 / 025, Loss: 0.962280

* Train accuracy / confusion: 57.38% / [[253, 90, 12], [84, 149, 37], [55, 63, 57]], 
* Val   accuracy / confusion: 52.88% / [[31, 12, 3], [7, 21, 7], [5, 15, 3]]

------------------------------ Epoch 402 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.847998
- Iter 024 / 025, Loss: 0.953935

* Train accuracy / confusion: 60.62% / [[264, 83, 11], [87, 147, 30], [48, 56, 74]], 
* Val   accuracy / confusion: 50.96% / [[33, 12, 1], [7, 19, 9], [4, 18, 1]]

------------------------------ Epoch 403 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.857956
- Iter 024 / 025, Loss: 0.802505

* Train accuracy / confusion: 59.75% / [[252, 93, 11], [85, 149, 33], [53, 47, 77]], 
* Val   accuracy / confusion: 52.88% / [[31, 12, 3], [6, 22, 7], [4, 17, 2]]

------------------------------ Epoch 404 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.975837
- Iter 024 / 025, Loss: 0.794182

* Train accuracy / confusion: 57.88% / [[251, 92, 11], [94, 142, 35], [48, 57, 70]], 
* Val   accuracy / confusion: 53.85% / [[30, 13, 3], [4, 23, 8], [4, 16, 3]]

------------------------------ Epoch 405 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.965781
- Iter 024 / 025, Loss: 0.950189

* Train accuracy / confusion: 58.12% / [[259, 89, 9], [93, 143, 31], [49, 64, 63]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [6, 22, 7], [4, 16, 3]]

------------------------------ Epoch 406 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.885381
- Iter 024 / 025, Loss: 0.926166

* Train accuracy / confusion: 57.88% / [[249, 95, 12], [88, 146, 36], [50, 56, 68]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [7, 22, 6], [5, 15, 3]]

------------------------------ Epoch 407 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.879612
- Iter 024 / 025, Loss: 0.984220

* Train accuracy / confusion: 59.00% / [[252, 91, 12], [91, 144, 36], [47, 51, 76]], 
* Val   accuracy / confusion: 51.92% / [[30, 12, 4], [8, 21, 6], [4, 16, 3]]

------------------------------ Epoch 408 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.963650
- Iter 024 / 025, Loss: 0.910773

* Train accuracy / confusion: 56.50% / [[250, 90, 14], [97, 136, 38], [55, 54, 66]], 
* Val   accuracy / confusion: 52.88% / [[31, 11, 4], [6, 22, 7], [7, 14, 2]]

------------------------------ Epoch 409 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.784830
- Iter 024 / 025, Loss: 0.845140

* Train accuracy / confusion: 58.75% / [[257, 88, 13], [94, 143, 30], [45, 60, 70]], 
* Val   accuracy / confusion: 57.69% / [[32, 10, 4], [6, 23, 6], [4, 14, 5]]

------------------------------ Epoch 410 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.010130
- Iter 024 / 025, Loss: 0.920155

* Train accuracy / confusion: 59.12% / [[257, 85, 12], [92, 147, 31], [49, 58, 69]], 
* Val   accuracy / confusion: 56.73% / [[31, 11, 4], [5, 25, 5], [3, 17, 3]]

------------------------------ Epoch 411 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.827985
- Iter 024 / 025, Loss: 0.782604

* Train accuracy / confusion: 59.62% / [[253, 92, 8], [90, 152, 28], [49, 56, 72]], 
* Val   accuracy / confusion: 54.81% / [[31, 11, 4], [5, 24, 6], [5, 16, 2]]

------------------------------ Epoch 412 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.795141
- Iter 024 / 025, Loss: 0.837741

* Train accuracy / confusion: 57.12% / [[249, 96, 9], [92, 144, 34], [47, 65, 64]], 
* Val   accuracy / confusion: 52.88% / [[30, 13, 3], [5, 23, 7], [4, 17, 2]]

------------------------------ Epoch 413 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.881275
- Iter 024 / 025, Loss: 0.755398

* Train accuracy / confusion: 58.88% / [[255, 92, 13], [94, 142, 29], [52, 49, 74]], 
* Val   accuracy / confusion: 55.77% / [[30, 13, 3], [6, 23, 6], [4, 14, 5]]

------------------------------ Epoch 414 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.895957
- Iter 024 / 025, Loss: 0.786119

* Train accuracy / confusion: 60.50% / [[254, 89, 14], [83, 159, 28], [47, 55, 71]], 
* Val   accuracy / confusion: 53.85% / [[29, 15, 2], [6, 22, 7], [5, 13, 5]]

------------------------------ Epoch 415 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.927931
- Iter 024 / 025, Loss: 0.870870

* Train accuracy / confusion: 58.50% / [[253, 89, 13], [92, 150, 28], [44, 66, 65]], 
* Val   accuracy / confusion: 54.81% / [[32, 11, 3], [8, 23, 4], [8, 13, 2]]

------------------------------ Epoch 416 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.037131
- Iter 024 / 025, Loss: 0.795516

* Train accuracy / confusion: 60.38% / [[258, 94, 9], [86, 153, 29], [49, 50, 72]], 
* Val   accuracy / confusion: 54.81% / [[30, 12, 4], [6, 24, 5], [4, 16, 3]]

------------------------------ Epoch 417 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.817482
- Iter 024 / 025, Loss: 0.864125

* Train accuracy / confusion: 59.38% / [[255, 87, 15], [91, 152, 29], [49, 54, 68]], 
* Val   accuracy / confusion: 55.77% / [[31, 11, 4], [6, 24, 5], [5, 15, 3]]

------------------------------ Epoch 418 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.800407
- Iter 024 / 025, Loss: 0.989258

* Train accuracy / confusion: 58.38% / [[254, 85, 13], [90, 148, 33], [50, 62, 65]], 
* Val   accuracy / confusion: 56.73% / [[31, 12, 3], [6, 23, 6], [4, 14, 5]]

------------------------------ Epoch 419 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.875881
- Iter 024 / 025, Loss: 0.764476

* Train accuracy / confusion: 59.25% / [[253, 95, 9], [90, 150, 29], [47, 56, 71]], 
* Val   accuracy / confusion: 51.92% / [[32, 11, 3], [9, 20, 6], [4, 17, 2]]

------------------------------ Epoch 420 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.900102
- Iter 024 / 025, Loss: 0.857047

* Train accuracy / confusion: 56.75% / [[246, 94, 12], [94, 141, 33], [52, 61, 67]], 
* Val   accuracy / confusion: 50.96% / [[32, 10, 4], [6, 20, 9], [4, 18, 1]]

------------------------------ Epoch 421 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.812668
- Iter 024 / 025, Loss: 0.755561

* Train accuracy / confusion: 59.75% / [[252, 88, 13], [92, 146, 31], [46, 52, 80]], 
* Val   accuracy / confusion: 51.92% / [[30, 11, 5], [6, 21, 8], [5, 15, 3]]

------------------------------ Epoch 422 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.968471
- Iter 024 / 025, Loss: 0.821982

* Train accuracy / confusion: 58.75% / [[247, 98, 11], [92, 151, 26], [49, 54, 72]], 
* Val   accuracy / confusion: 52.88% / [[31, 12, 3], [6, 23, 6], [4, 18, 1]]

------------------------------ Epoch 423 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.992409
- Iter 024 / 025, Loss: 0.874062

* Train accuracy / confusion: 58.75% / [[255, 93, 12], [84, 147, 34], [50, 57, 68]], 
* Val   accuracy / confusion: 54.81% / [[31, 12, 3], [7, 23, 5], [4, 16, 3]]

------------------------------ Epoch 424 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.142277
- Iter 024 / 025, Loss: 0.703867

* Train accuracy / confusion: 61.25% / [[256, 84, 14], [84, 155, 28], [52, 48, 79]], 
* Val   accuracy / confusion: 53.85% / [[29, 12, 5], [5, 24, 6], [5, 15, 3]]

------------------------------ Epoch 425 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.873913
- Iter 024 / 025, Loss: 0.935930

* Train accuracy / confusion: 59.25% / [[259, 84, 14], [81, 147, 39], [52, 56, 68]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [7, 22, 6], [4, 16, 3]]

------------------------------ Epoch 426 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.764438
- Iter 024 / 025, Loss: 0.813869

* Train accuracy / confusion: 57.88% / [[258, 85, 14], [98, 137, 34], [47, 59, 68]], 
* Val   accuracy / confusion: 57.69% / [[32, 10, 4], [6, 25, 4], [5, 15, 3]]

------------------------------ Epoch 427 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.852731
- Iter 024 / 025, Loss: 1.104433

* Train accuracy / confusion: 58.12% / [[249, 93, 13], [93, 149, 28], [51, 57, 67]], 
* Val   accuracy / confusion: 52.88% / [[32, 9, 5], [8, 21, 6], [6, 15, 2]]

------------------------------ Epoch 428 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.960833
- Iter 024 / 025, Loss: 0.971835

* Train accuracy / confusion: 60.12% / [[256, 81, 15], [91, 152, 25], [51, 56, 73]], 
* Val   accuracy / confusion: 53.85% / [[30, 14, 2], [8, 22, 5], [5, 14, 4]]

------------------------------ Epoch 429 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.883384
- Iter 024 / 025, Loss: 0.934601

* Train accuracy / confusion: 57.75% / [[247, 96, 13], [90, 148, 29], [53, 57, 67]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [6, 22, 7], [5, 15, 3]]

------------------------------ Epoch 430 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.970105
- Iter 024 / 025, Loss: 0.849251

* Train accuracy / confusion: 61.12% / [[264, 83, 10], [84, 150, 32], [47, 55, 75]], 
* Val   accuracy / confusion: 53.85% / [[32, 10, 4], [7, 21, 7], [3, 17, 3]]

------------------------------ Epoch 431 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.841604
- Iter 024 / 025, Loss: 0.799596

* Train accuracy / confusion: 59.75% / [[254, 87, 17], [81, 152, 35], [49, 53, 72]], 
* Val   accuracy / confusion: 52.88% / [[30, 12, 4], [5, 24, 6], [5, 17, 1]]

------------------------------ Epoch 432 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.960850
- Iter 024 / 025, Loss: 0.815868

* Train accuracy / confusion: 60.38% / [[262, 85, 8], [87, 152, 29], [49, 59, 69]], 
* Val   accuracy / confusion: 49.04% / [[32, 11, 3], [8, 19, 8], [5, 18, 0]]

------------------------------ Epoch 433 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.882575
- Iter 024 / 025, Loss: 0.979918

* Train accuracy / confusion: 60.25% / [[260, 84, 12], [88, 147, 31], [46, 57, 75]], 
* Val   accuracy / confusion: 55.77% / [[31, 11, 4], [6, 24, 5], [4, 16, 3]]

------------------------------ Epoch 434 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.908973
- Iter 024 / 025, Loss: 0.989088

* Train accuracy / confusion: 57.25% / [[249, 97, 14], [98, 138, 30], [51, 52, 71]], 
* Val   accuracy / confusion: 59.62% / [[31, 10, 5], [5, 26, 4], [4, 14, 5]]

------------------------------ Epoch 435 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.899305
- Iter 024 / 025, Loss: 0.790961

* Train accuracy / confusion: 57.50% / [[254, 94, 12], [91, 142, 33], [48, 62, 64]], 
* Val   accuracy / confusion: 53.85% / [[29, 14, 3], [6, 24, 5], [5, 15, 3]]

------------------------------ Epoch 436 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.975187
- Iter 024 / 025, Loss: 0.881344

* Train accuracy / confusion: 56.62% / [[250, 92, 13], [97, 136, 38], [48, 59, 67]], 
* Val   accuracy / confusion: 51.92% / [[30, 13, 3], [5, 22, 8], [4, 17, 2]]

------------------------------ Epoch 437 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.879741
- Iter 024 / 025, Loss: 0.944697

* Train accuracy / confusion: 57.88% / [[253, 90, 12], [98, 139, 32], [46, 59, 71]], 
* Val   accuracy / confusion: 55.77% / [[31, 12, 3], [6, 23, 6], [5, 14, 4]]

------------------------------ Epoch 438 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.925199
- Iter 024 / 025, Loss: 0.901726

* Train accuracy / confusion: 57.12% / [[252, 90, 16], [95, 139, 35], [53, 54, 66]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 439 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.007716
- Iter 024 / 025, Loss: 0.812088

* Train accuracy / confusion: 59.12% / [[248, 94, 12], [97, 151, 20], [54, 50, 74]], 
* Val   accuracy / confusion: 56.73% / [[33, 9, 4], [7, 23, 5], [5, 15, 3]]

------------------------------ Epoch 440 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.924333
- Iter 024 / 025, Loss: 0.762458

* Train accuracy / confusion: 57.38% / [[254, 85, 15], [95, 136, 39], [53, 54, 69]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [7, 21, 7], [3, 18, 2]]

------------------------------ Epoch 441 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.840354
- Iter 024 / 025, Loss: 0.835302

* Train accuracy / confusion: 57.12% / [[248, 96, 13], [97, 138, 31], [44, 62, 71]], 
* Val   accuracy / confusion: 52.88% / [[30, 12, 4], [6, 23, 6], [5, 16, 2]]

------------------------------ Epoch 442 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.892061
- Iter 024 / 025, Loss: 0.898887

* Train accuracy / confusion: 58.12% / [[249, 92, 15], [90, 150, 32], [48, 58, 66]], 
* Val   accuracy / confusion: 51.92% / [[31, 11, 4], [6, 22, 7], [5, 17, 1]]

------------------------------ Epoch 443 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.137713
- Iter 024 / 025, Loss: 0.765296

* Train accuracy / confusion: 59.50% / [[254, 86, 15], [84, 157, 27], [51, 61, 65]], 
* Val   accuracy / confusion: 54.81% / [[32, 9, 5], [6, 21, 8], [4, 15, 4]]

------------------------------ Epoch 444 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.916644
- Iter 024 / 025, Loss: 0.805161

* Train accuracy / confusion: 59.12% / [[253, 86, 14], [88, 155, 27], [52, 60, 65]], 
* Val   accuracy / confusion: 53.85% / [[31, 13, 2], [5, 23, 7], [4, 17, 2]]

------------------------------ Epoch 445 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.009141
- Iter 024 / 025, Loss: 0.851895

* Train accuracy / confusion: 58.75% / [[250, 100, 10], [86, 152, 28], [50, 56, 68]], 
* Val   accuracy / confusion: 57.69% / [[32, 10, 4], [4, 26, 5], [4, 17, 2]]

------------------------------ Epoch 446 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.883751
- Iter 024 / 025, Loss: 0.976812

* Train accuracy / confusion: 59.50% / [[259, 87, 9], [93, 147, 29], [47, 59, 70]], 
* Val   accuracy / confusion: 54.81% / [[30, 13, 3], [5, 25, 5], [4, 17, 2]]

------------------------------ Epoch 447 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.716066
- Iter 024 / 025, Loss: 0.908554

* Train accuracy / confusion: 60.38% / [[266, 83, 11], [85, 146, 30], [49, 59, 71]], 
* Val   accuracy / confusion: 58.65% / [[32, 12, 2], [5, 23, 7], [4, 13, 6]]

------------------------------ Epoch 448 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.748931
- Iter 024 / 025, Loss: 0.878140

* Train accuracy / confusion: 57.88% / [[250, 94, 13], [91, 144, 34], [48, 57, 69]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [7, 22, 6], [4, 16, 3]]

------------------------------ Epoch 449 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.924016
- Iter 024 / 025, Loss: 0.865980

* Train accuracy / confusion: 60.75% / [[257, 91, 10], [82, 155, 26], [54, 51, 74]], 
* Val   accuracy / confusion: 54.81% / [[31, 13, 2], [6, 21, 8], [4, 14, 5]]

------------------------------ Epoch 450 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.984125
- Iter 024 / 025, Loss: 0.795463

* Train accuracy / confusion: 58.50% / [[253, 89, 14], [84, 152, 32], [45, 68, 63]], 
* Val   accuracy / confusion: 55.77% / [[31, 13, 2], [4, 24, 7], [5, 15, 3]]

------------------------------ Epoch 451 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.858685
- Iter 024 / 025, Loss: 0.789059

* Train accuracy / confusion: 58.50% / [[256, 84, 15], [94, 142, 34], [55, 50, 70]], 
* Val   accuracy / confusion: 50.96% / [[29, 14, 3], [5, 23, 7], [5, 17, 1]]

------------------------------ Epoch 452 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.114137
- Iter 024 / 025, Loss: 0.701128

* Train accuracy / confusion: 58.12% / [[254, 90, 12], [92, 139, 36], [50, 55, 72]], 
* Val   accuracy / confusion: 53.85% / [[32, 12, 2], [6, 23, 6], [4, 18, 1]]

------------------------------ Epoch 453 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.755754
- Iter 024 / 025, Loss: 0.906308

* Train accuracy / confusion: 59.00% / [[252, 96, 13], [90, 148, 27], [49, 53, 72]], 
* Val   accuracy / confusion: 51.92% / [[28, 14, 4], [5, 23, 7], [4, 16, 3]]

------------------------------ Epoch 454 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.952055
- Iter 024 / 025, Loss: 0.940672

* Train accuracy / confusion: 59.62% / [[260, 83, 11], [89, 147, 33], [49, 58, 70]], 
* Val   accuracy / confusion: 55.77% / [[32, 12, 2], [6, 23, 6], [4, 16, 3]]

------------------------------ Epoch 455 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.917698
- Iter 024 / 025, Loss: 0.839949

* Train accuracy / confusion: 58.38% / [[254, 89, 14], [95, 142, 30], [52, 53, 71]], 
* Val   accuracy / confusion: 52.88% / [[32, 11, 3], [7, 22, 6], [4, 18, 1]]

------------------------------ Epoch 456 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.941065
- Iter 024 / 025, Loss: 0.936424

* Train accuracy / confusion: 58.75% / [[261, 86, 10], [91, 140, 35], [47, 61, 69]], 
* Val   accuracy / confusion: 57.69% / [[33, 11, 2], [6, 26, 3], [5, 17, 1]]

------------------------------ Epoch 457 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.820043
- Iter 024 / 025, Loss: 0.978241

* Train accuracy / confusion: 58.25% / [[256, 89, 11], [87, 143, 37], [49, 61, 67]], 
* Val   accuracy / confusion: 54.81% / [[31, 14, 1], [7, 23, 5], [4, 16, 3]]

------------------------------ Epoch 458 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.952686
- Iter 024 / 025, Loss: 0.812112

* Train accuracy / confusion: 59.38% / [[260, 88, 9], [90, 146, 34], [50, 54, 69]], 
* Val   accuracy / confusion: 51.92% / [[31, 10, 5], [8, 21, 6], [4, 17, 2]]

------------------------------ Epoch 459 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.805708
- Iter 024 / 025, Loss: 0.860851

* Train accuracy / confusion: 59.25% / [[252, 90, 16], [90, 149, 29], [47, 54, 73]], 
* Val   accuracy / confusion: 55.77% / [[31, 13, 2], [4, 24, 7], [5, 15, 3]]

------------------------------ Epoch 460 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.891744
- Iter 024 / 025, Loss: 0.926033

* Train accuracy / confusion: 59.38% / [[251, 91, 12], [93, 150, 27], [48, 54, 74]], 
* Val   accuracy / confusion: 55.77% / [[33, 11, 2], [6, 22, 7], [6, 14, 3]]

------------------------------ Epoch 461 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.911252
- Iter 024 / 025, Loss: 0.919506

* Train accuracy / confusion: 59.00% / [[252, 92, 14], [86, 151, 30], [52, 54, 69]], 
* Val   accuracy / confusion: 52.88% / [[31, 13, 2], [7, 22, 6], [4, 17, 2]]

------------------------------ Epoch 462 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.978776
- Iter 024 / 025, Loss: 0.873122

* Train accuracy / confusion: 58.50% / [[247, 96, 7], [95, 148, 30], [51, 53, 73]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [7, 24, 4], [5, 16, 2]]

------------------------------ Epoch 463 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.911577
- Iter 024 / 025, Loss: 0.895744

* Train accuracy / confusion: 59.25% / [[259, 84, 11], [89, 143, 36], [48, 58, 72]], 
* Val   accuracy / confusion: 53.85% / [[30, 13, 3], [8, 23, 4], [5, 15, 3]]

------------------------------ Epoch 464 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.922959
- Iter 024 / 025, Loss: 1.089077

* Train accuracy / confusion: 59.75% / [[250, 92, 10], [87, 154, 32], [47, 54, 74]], 
* Val   accuracy / confusion: 56.73% / [[32, 10, 4], [7, 23, 5], [5, 14, 4]]

------------------------------ Epoch 465 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.813472
- Iter 024 / 025, Loss: 1.119783

* Train accuracy / confusion: 57.75% / [[243, 96, 12], [96, 147, 31], [53, 50, 72]], 
* Val   accuracy / confusion: 51.92% / [[32, 9, 5], [7, 21, 7], [5, 17, 1]]

------------------------------ Epoch 466 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.943621
- Iter 024 / 025, Loss: 0.964736

* Train accuracy / confusion: 58.88% / [[254, 94, 11], [83, 148, 32], [49, 60, 69]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [6, 22, 7], [4, 17, 2]]

------------------------------ Epoch 467 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.820791
- Iter 024 / 025, Loss: 0.791195

* Train accuracy / confusion: 60.88% / [[257, 84, 15], [89, 152, 27], [50, 48, 78]], 
* Val   accuracy / confusion: 58.65% / [[32, 11, 3], [5, 27, 3], [4, 17, 2]]

------------------------------ Epoch 468 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.921033
- Iter 024 / 025, Loss: 0.795278

* Train accuracy / confusion: 61.38% / [[258, 77, 15], [82, 158, 33], [45, 57, 75]], 
* Val   accuracy / confusion: 54.81% / [[31, 12, 3], [5, 23, 7], [5, 15, 3]]

------------------------------ Epoch 469 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.931865
- Iter 024 / 025, Loss: 0.919311

* Train accuracy / confusion: 60.62% / [[258, 87, 12], [84, 150, 30], [47, 55, 77]], 
* Val   accuracy / confusion: 55.77% / [[31, 14, 1], [5, 25, 5], [4, 17, 2]]

------------------------------ Epoch 470 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.765280
- Iter 024 / 025, Loss: 0.966397

* Train accuracy / confusion: 58.38% / [[244, 98, 11], [94, 150, 25], [44, 61, 73]], 
* Val   accuracy / confusion: 58.65% / [[34, 8, 4], [6, 24, 5], [6, 14, 3]]

------------------------------ Epoch 471 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.804210
- Iter 024 / 025, Loss: 0.907670

* Train accuracy / confusion: 57.25% / [[249, 94, 11], [93, 140, 38], [53, 53, 69]], 
* Val   accuracy / confusion: 50.00% / [[31, 11, 4], [9, 20, 6], [4, 18, 1]]

------------------------------ Epoch 472 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.819742
- Iter 024 / 025, Loss: 1.112169

* Train accuracy / confusion: 58.75% / [[255, 89, 17], [86, 150, 28], [49, 61, 65]], 
* Val   accuracy / confusion: 55.77% / [[30, 14, 2], [4, 26, 5], [5, 16, 2]]

------------------------------ Epoch 473 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.790516
- Iter 024 / 025, Loss: 0.762519

* Train accuracy / confusion: 59.25% / [[251, 96, 8], [82, 157, 31], [52, 57, 66]], 
* Val   accuracy / confusion: 57.69% / [[30, 13, 3], [6, 26, 3], [4, 15, 4]]

------------------------------ Epoch 474 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.145437
- Iter 024 / 025, Loss: 0.888985

* Train accuracy / confusion: 58.62% / [[256, 94, 10], [84, 148, 33], [50, 60, 65]], 
* Val   accuracy / confusion: 57.69% / [[31, 14, 1], [6, 26, 3], [4, 16, 3]]

------------------------------ Epoch 475 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.929320
- Iter 024 / 025, Loss: 0.890539

* Train accuracy / confusion: 61.00% / [[261, 88, 10], [84, 154, 27], [46, 57, 73]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [5, 23, 7], [4, 16, 3]]

------------------------------ Epoch 476 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.950889
- Iter 024 / 025, Loss: 0.837150

* Train accuracy / confusion: 58.50% / [[249, 101, 8], [87, 148, 33], [43, 60, 71]], 
* Val   accuracy / confusion: 56.73% / [[31, 11, 4], [4, 25, 6], [4, 16, 3]]

------------------------------ Epoch 477 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.898203
- Iter 024 / 025, Loss: 0.900996

* Train accuracy / confusion: 60.38% / [[251, 93, 10], [86, 159, 25], [45, 58, 73]], 
* Val   accuracy / confusion: 51.92% / [[30, 13, 3], [7, 21, 7], [5, 15, 3]]

------------------------------ Epoch 478 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.924463
- Iter 024 / 025, Loss: 0.850407

* Train accuracy / confusion: 58.50% / [[255, 90, 10], [95, 144, 33], [47, 57, 69]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [6, 23, 6], [5, 16, 2]]

------------------------------ Epoch 479 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.942904
- Iter 024 / 025, Loss: 0.832474

* Train accuracy / confusion: 59.88% / [[252, 99, 8], [89, 151, 26], [48, 51, 76]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [7, 24, 4], [6, 15, 2]]

------------------------------ Epoch 480 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.075601
- Iter 024 / 025, Loss: 0.950052

* Train accuracy / confusion: 58.75% / [[254, 87, 16], [87, 149, 34], [47, 59, 67]], 
* Val   accuracy / confusion: 55.77% / [[33, 8, 5], [7, 20, 8], [4, 14, 5]]

------------------------------ Epoch 481 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.960093
- Iter 024 / 025, Loss: 0.870436

* Train accuracy / confusion: 59.25% / [[252, 91, 16], [92, 149, 25], [47, 55, 73]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [9, 21, 5], [4, 15, 4]]

------------------------------ Epoch 482 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.966175
- Iter 024 / 025, Loss: 0.874648

* Train accuracy / confusion: 58.88% / [[250, 95, 11], [90, 147, 32], [50, 51, 74]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [5, 23, 7], [4, 17, 2]]

------------------------------ Epoch 483 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.822520
- Iter 024 / 025, Loss: 0.767018

* Train accuracy / confusion: 60.25% / [[256, 92, 8], [80, 156, 30], [51, 57, 70]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [5, 22, 8], [3, 17, 3]]

------------------------------ Epoch 484 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.883250
- Iter 024 / 025, Loss: 0.731677

* Train accuracy / confusion: 61.38% / [[255, 91, 8], [85, 156, 29], [49, 47, 80]], 
* Val   accuracy / confusion: 56.73% / [[33, 8, 5], [5, 23, 7], [5, 15, 3]]

------------------------------ Epoch 485 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.860419
- Iter 024 / 025, Loss: 0.849810

* Train accuracy / confusion: 60.12% / [[252, 91, 13], [91, 154, 27], [51, 46, 75]], 
* Val   accuracy / confusion: 50.96% / [[30, 13, 3], [7, 20, 8], [5, 15, 3]]

------------------------------ Epoch 486 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.769623
- Iter 024 / 025, Loss: 0.992600

* Train accuracy / confusion: 56.62% / [[246, 100, 12], [92, 144, 26], [54, 63, 63]], 
* Val   accuracy / confusion: 56.73% / [[31, 12, 3], [4, 25, 6], [3, 17, 3]]

------------------------------ Epoch 487 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.939682
- Iter 024 / 025, Loss: 0.911542

* Train accuracy / confusion: 60.62% / [[259, 82, 14], [87, 153, 28], [48, 56, 73]], 
* Val   accuracy / confusion: 58.65% / [[34, 10, 2], [6, 25, 4], [4, 17, 2]]

------------------------------ Epoch 488 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.063269
- Iter 024 / 025, Loss: 1.064679

* Train accuracy / confusion: 60.00% / [[256, 91, 10], [84, 151, 32], [49, 54, 73]], 
* Val   accuracy / confusion: 50.00% / [[30, 12, 4], [6, 20, 9], [5, 16, 2]]

------------------------------ Epoch 489 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.887674
- Iter 024 / 025, Loss: 0.917404

* Train accuracy / confusion: 59.25% / [[255, 97, 10], [88, 149, 31], [47, 53, 70]], 
* Val   accuracy / confusion: 56.73% / [[32, 10, 4], [5, 25, 5], [4, 17, 2]]

------------------------------ Epoch 490 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.887458
- Iter 024 / 025, Loss: 0.835628

* Train accuracy / confusion: 59.88% / [[258, 84, 15], [92, 149, 29], [50, 51, 72]], 
* Val   accuracy / confusion: 54.81% / [[31, 10, 5], [5, 24, 6], [6, 15, 2]]

------------------------------ Epoch 491 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.810182
- Iter 024 / 025, Loss: 0.946682

* Train accuracy / confusion: 58.75% / [[260, 83, 15], [89, 144, 36], [48, 59, 66]], 
* Val   accuracy / confusion: 52.88% / [[31, 11, 4], [6, 21, 8], [4, 16, 3]]

------------------------------ Epoch 492 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 1.060190
- Iter 024 / 025, Loss: 0.839486

* Train accuracy / confusion: 59.75% / [[256, 86, 14], [85, 156, 26], [46, 65, 66]], 
* Val   accuracy / confusion: 53.85% / [[32, 13, 1], [7, 22, 6], [6, 15, 2]]

------------------------------ Epoch 493 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.801766
- Iter 024 / 025, Loss: 0.918354

* Train accuracy / confusion: 59.38% / [[255, 88, 13], [91, 150, 26], [49, 58, 70]], 
* Val   accuracy / confusion: 56.73% / [[32, 12, 2], [6, 25, 4], [4, 17, 2]]

------------------------------ Epoch 494 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.944708
- Iter 024 / 025, Loss: 0.780328

* Train accuracy / confusion: 59.62% / [[256, 87, 10], [82, 155, 34], [49, 61, 66]], 
* Val   accuracy / confusion: 55.77% / [[31, 12, 3], [6, 23, 6], [5, 14, 4]]

------------------------------ Epoch 495 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.765899
- Iter 024 / 025, Loss: 0.824835

* Train accuracy / confusion: 59.62% / [[249, 99, 10], [93, 152, 21], [46, 54, 76]], 
* Val   accuracy / confusion: 57.69% / [[32, 12, 2], [3, 26, 6], [3, 18, 2]]

------------------------------ Epoch 496 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.830718
- Iter 024 / 025, Loss: 0.787379

* Train accuracy / confusion: 60.12% / [[260, 91, 7], [81, 151, 37], [44, 59, 70]], 
* Val   accuracy / confusion: 54.81% / [[33, 10, 3], [6, 22, 7], [3, 18, 2]]

------------------------------ Epoch 497 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.876788
- Iter 024 / 025, Loss: 0.694323

* Train accuracy / confusion: 60.00% / [[266, 82, 12], [89, 148, 29], [48, 60, 66]], 
* Val   accuracy / confusion: 55.77% / [[32, 8, 6], [7, 23, 5], [5, 15, 3]]

------------------------------ Epoch 498 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.857617
- Iter 024 / 025, Loss: 0.970759

* Train accuracy / confusion: 59.62% / [[255, 87, 12], [92, 155, 24], [47, 61, 67]], 
* Val   accuracy / confusion: 57.69% / [[31, 11, 4], [5, 26, 4], [4, 16, 3]]

------------------------------ Epoch 499 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.889189
- Iter 024 / 025, Loss: 0.805312

* Train accuracy / confusion: 59.00% / [[247, 93, 16], [88, 157, 26], [54, 51, 68]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [6, 22, 7], [4, 16, 3]]

------------------------------ Epoch 500 / 500, Learning rate: 3.28e-04 ------------------------------
- Iter 012 / 025, Loss: 0.952523
- Iter 024 / 025, Loss: 0.978863

* Train accuracy / confusion: 59.38% / [[254, 88, 12], [84, 151, 33], [45, 63, 70]], 
* Val   accuracy / confusion: 51.92% / [[31, 12, 3], [8, 21, 6], [5, 16, 2]]

**************************************** Training Ends ****************************************
- Test accuracy: 57.54%

- Confusion matrix:
 [[922 402  56]
 [192 656 172]
 [ 96 394 200]]

- Debug table:
{   '00941': {'GT': 0, 'Acc': ' 73.33%', 'Pred': [22, 8, 0]},
    '00408': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00871': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00852': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00816': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00672': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00414': {'GT': 2, 'Acc': ' 86.67%', 'Pred': [0, 4, 26]},
    '00767': {'GT': 1, 'Acc': ' 60.00%', 'Pred': [12, 18, 0]},
    '01148': {'GT': 1, 'Acc': ' 46.67%', 'Pred': [16, 14, 0]},
    '00743': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00305': {'GT': 1, 'Acc': ' 86.67%', 'Pred': [0, 26, 4]},
    '01305': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00455': {'GT': 1, 'Acc': ' 46.67%', 'Pred': [0, 14, 16]},
    '01140': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00649': {'GT': 2, 'Acc': ' 56.67%', 'Pred': [0, 13, 17]},
    '01068': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00176': {'GT': 0, 'Acc': '  3.33%', 'Pred': [1, 29, 0]},
    '00861': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00302': {'GT': 1, 'Acc': ' 70.00%', 'Pred': [0, 21, 9]},
    '00756': {'GT': 1, 'Acc': ' 46.67%', 'Pred': [16, 14, 0]},
    '00706': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01218': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '01008': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00914': {'GT': 0, 'Acc': ' 36.67%', 'Pred': [11, 19, 0]},
    '00680': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00736': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00417': {'GT': 2, 'Acc': '  6.67%', 'Pred': [0, 28, 2]},
    '01184': {'GT': 2, 'Acc': ' 43.33%', 'Pred': [0, 17, 13]},
    '00949': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '01268': {'GT': 1, 'Acc': ' 26.67%', 'Pred': [22, 8, 0]},
    '00441': {'GT': 0, 'Acc': ' 76.67%', 'Pred': [23, 7, 0]},
    '00731': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00806': {'GT': 2, 'Acc': ' 30.00%', 'Pred': [4, 17, 9]},
    '00192': {'GT': 0, 'Acc': ' 26.67%', 'Pred': [8, 22, 0]},
    '01075': {'GT': 2, 'Acc': ' 36.67%', 'Pred': [0, 19, 11]},
    '00058': {'GT': 0, 'Acc': ' 30.00%', 'Pred': [9, 11, 10]},
    '01126': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00924': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01070': {'GT': 0, 'Acc': ' 30.00%', 'Pred': [9, 21, 0]},
    '00901': {'GT': 0, 'Acc': '  6.67%', 'Pred': [2, 28, 0]},
    '01079': {'GT': 1, 'Acc': ' 73.33%', 'Pred': [8, 22, 0]},
    '00614': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00062': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00405': {'GT': 2, 'Acc': '100.00%', 'Pred': [0, 0, 30]},
    '01307': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00918': {'GT': 0, 'Acc': '  3.33%', 'Pred': [1, 29, 0]},
    '00299': {'GT': 0, 'Acc': '  3.33%', 'Pred': [1, 28, 1]},
    '01046': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01106': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00822': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01037': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00599': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00517': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00832': {'GT': 2, 'Acc': ' 10.00%', 'Pred': [0, 27, 3]},
    '00498': {'GT': 1, 'Acc': '  0.00%', 'Pred': [30, 0, 0]},
    '00280': {'GT': 2, 'Acc': '100.00%', 'Pred': [0, 0, 30]},
    '00623': {'GT': 2, 'Acc': '  0.00%', 'Pred': [30, 0, 0]},
    '01038': {'GT': 1, 'Acc': ' 36.67%', 'Pred': [0, 11, 19]},
    '00534': {'GT': 0, 'Acc': '  3.33%', 'Pred': [1, 29, 0]},
    '01199': {'GT': 2, 'Acc': ' 60.00%', 'Pred': [0, 12, 18]},
    '00886': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 15, 15]},
    '00698': {'GT': 1, 'Acc': ' 86.67%', 'Pred': [0, 26, 4]},
    '00509': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00389': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '01208': {'GT': 2, 'Acc': '  0.00%', 'Pred': [6, 24, 0]},
    '00004': {'GT': 2, 'Acc': ' 53.33%', 'Pred': [0, 14, 16]},
    '01039': {'GT': 1, 'Acc': ' 73.33%', 'Pred': [0, 22, 8]},
    '01086': {'GT': 2, 'Acc': '  0.00%', 'Pred': [13, 17, 0]},
    '00174': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00226': {'GT': 2, 'Acc': ' 56.67%', 'Pred': [0, 13, 17]},
    '00794': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00289': {'GT': 2, 'Acc': ' 23.33%', 'Pred': [0, 23, 7]},
    '01027': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00855': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00385': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 0, 30]},
    '00022': {'GT': 1, 'Acc': ' 80.00%', 'Pred': [0, 24, 6]},
    '00220': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00635': {'GT': 2, 'Acc': '  0.00%', 'Pred': [30, 0, 0]},
    '00122': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00505': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01014': {'GT': 1, 'Acc': '  0.00%', 'Pred': [30, 0, 0]},
    '00407': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '01300': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00301': {'GT': 1, 'Acc': ' 73.33%', 'Pred': [0, 22, 8]},
    '01349': {'GT': 1, 'Acc': ' 96.67%', 'Pred': [0, 29, 1]},
    '00005': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '01169': {'GT': 0, 'Acc': '  3.33%', 'Pred': [1, 29, 0]},
    '00124': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00243': {'GT': 1, 'Acc': ' 26.67%', 'Pred': [22, 8, 0]},
    '01240': {'GT': 1, 'Acc': '  6.67%', 'Pred': [0, 2, 28]},
    '00589': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01172': {'GT': 1, 'Acc': ' 70.00%', 'Pred': [0, 21, 9]},
    '01156': {'GT': 1, 'Acc': '  0.00%', 'Pred': [30, 0, 0]},
    '00329': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01333': {'GT': 0, 'Acc': ' 86.67%', 'Pred': [26, 4, 0]},
    '01364': {'GT': 1, 'Acc': '  0.00%', 'Pred': [0, 0, 30]},
    '00415': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00603': {'GT': 2, 'Acc': '  3.33%', 'Pred': [13, 16, 1]},
    '00096': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01250': {'GT': 1, 'Acc': ' 80.00%', 'Pred': [6, 24, 0]},
    '00125': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 3, 0]},
    '00447': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00820': {'GT': 1, 'Acc': '  0.00%', 'Pred': [0, 0, 30]}}

1D CNN: Siamese model4

class SiameseNet4(nn.Module):
    def __init__(self, n_input=20, n_output=3, n_channel=64,
                 use_age=True, final_pool='average'):
        super().__init__()
        
        if final_pool not in {'average', 'max'}:
            raise ValueError("final_pool must be set to one of ['average', 'max']")
        
        self.use_age = use_age
        
        self.conv1 = nn.Conv1d(n_input, n_channel, kernel_size=25, stride=2)
        self.bn1 = nn.BatchNorm1d(n_channel)
        self.pool1 = nn.MaxPool1d(2)
        
        self.conv2 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn2 = nn.BatchNorm1d(n_channel)
        self.pool2 = nn.MaxPool1d(2)
        
        self.conv3 = nn.Conv1d(n_channel, n_channel, kernel_size=7)
        self.bn3 = nn.BatchNorm1d(n_channel)
        self.pool3 = nn.MaxPool1d(2)
        
        self.conv4_1 = nn.Conv1d(n_channel, n_channel*2, kernel_size=9)
        self.bn4_1 = nn.BatchNorm1d(n_channel*2)

        self.conv4_2 = nn.Conv1d(n_channel, n_channel*2, kernel_size=9)
        self.bn4_2 = nn.BatchNorm1d(n_channel*2)

        self.conv5_1 = nn.Conv1d(n_channel*2, n_channel*2, kernel_size=7)
        self.bn5_1 = nn.BatchNorm1d(n_channel*2)

        self.conv5_2 = nn.Conv1d(n_channel*2, n_channel*2, kernel_size=7)
        self.bn5_2 = nn.BatchNorm1d(n_channel*2)
        
        if final_pool == 'average':
            self.final_pool = nn.AdaptiveAvgPool1d(1)
        elif final_pool == 'max':
            self.final_pool = nn.AdaptiveMaxPool1d(1)
            
        if self.use_age:        
            self.fc1 = nn.Linear(n_channel*4 + 1, n_channel)
        else:
            self.fc1 = nn.Linear(n_channel*4, n_channel)
            
        self.dropout = nn.Dropout(p=0.2)
        self.bnfc1 = nn.BatchNorm1d(n_channel)

        self.fc2 = nn.Linear(n_channel, n_channel//2)
        self.bnfc2 = nn.BatchNorm1d(n_channel//2)
        self.fc3 = nn.Linear(n_channel//2, n_output)
        
    def reset_weights(self):
        for m in self.modules():
            if hasattr(m, 'reset_parameters'):
                m.reset_parameters()

    def forward(self, x, age):
        # conv-bn-relu-pool 
        x = self.conv1(x)
        x = F.relu(self.bn1(x))
        x = self.pool1(x)

        x = self.conv2(x)
        x = F.relu(self.bn2(x))
        x = self.pool2(x)

        x = self.conv3(x)
        x = F.relu(self.bn3(x))
        x = self.pool3(x)
        
        L = x.shape[2]
        x1 = x[:, :, :L//2]
        x2 = x[:, :, L//2:]
        
        x1 = self.conv4_1(x1)
        x1 = F.relu(self.bn4_1(x1))

        x2 = self.conv4_2(x2)
        x2 = F.relu(self.bn4_2(x2))
        
        x1 = self.conv5_1(x1)
        x1 = F.relu(self.bn5_1(x1))

        x2 = self.conv5_2(x2)
        x2 = F.relu(self.bn5_2(x2))
                
        x1 = self.final_pool(x1).squeeze()
        x2 = self.final_pool(x2).squeeze()

        x = torch.cat((x1, x2), dim=1)
        if self.use_age:
            x = torch.cat((x, age.reshape(-1, 1)), dim=1)

        # fc-bn-dropout-relu-fc
        x = self.fc1(x)
        x = self.bnfc1(x)
        x = self.dropout(x)
        x = F.relu(x)
        
        x = self.fc2(x)
        x = self.bnfc2(x)
        x = F.relu(x)
        
        x = self.fc3(x)
        
        return x
        # return F.log_softmax(x, dim=1)

model = SiameseNet4(n_input=train_dataset[0]['signal'].shape[0], 
                    n_output=3, 
                    use_age=True, 
                    final_pool='max')

model = model.to(device, dtype=torch.float32)
print(model)
print()

# tensorboard visualization
visualize_network_tensorboard(model, 'SiameseNet_3')

# number of parameters
n = count_parameters(model)
print(f'The Number of parameters of the model: {n:,}')

SiameseNet4(
  (conv1): Conv1d(20, 64, kernel_size=(25,), stride=(2,))
  (bn1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool1): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv3): Conv1d(64, 64, kernel_size=(7,), stride=(1,))
  (bn3): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (pool3): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv4_1): Conv1d(64, 128, kernel_size=(9,), stride=(1,))
  (bn4_1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv4_2): Conv1d(64, 128, kernel_size=(9,), stride=(1,))
  (bn4_2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv5_1): Conv1d(128, 128, kernel_size=(7,), stride=(1,))
  (bn5_1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv5_2): Conv1d(128, 128, kernel_size=(7,), stride=(1,))
  (bn5_2): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (final_pool): AdaptiveMaxPool1d(output_size=1)
  (fc1): Linear(in_features=257, out_features=64, bias=True)
  (dropout): Dropout(p=0.2, inplace=False)
  (bnfc1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc2): Linear(in_features=64, out_features=32, bias=True)
  (bnfc2): BatchNorm1d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc3): Linear(in_features=32, out_features=3, bias=True)
)

The Number of parameters of the model: 487,171

record = learning_rate_search(model, 
                              min_log_lr=-5.0, 
                              max_log_lr=-1.0, 
                              trials=500, 
                              epochs=1)
draw_learning_rate_record(record)
best_log_lr = record[np.argmax(np.array([v for lr, v in record]))][0]

# best_log_lr = -2.0523682533344423
print('best_log_lr:', best_log_lr)

best_log_lr: -3.3652163701672517

# reduce the learning after [lr_schedule_step] epochs by a factor of 10
n_epoch = 500
lr_schedule_step = 200
log_interval = len(train_loader) // 2

loss_history = []
train_acc_history = []
val_acc_history = []

model.reset_weights()
optimizer = optim.AdamW(model.parameters(), lr=10 ** best_log_lr, weight_decay=0.0001)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=lr_schedule_step, gamma=0.1)

print(f'{"*"*40} Training Starts {"*"*40}')

for epoch in tqdm(range(1, n_epoch + 1)):
    print(f'{"-"*30} Epoch {epoch:03d} / {n_epoch:03d}, Learning rate: {optimizer.param_groups[-1]["lr"]:.2e} {"-"*30}')
    
    # train
    loss, train_accuracy, train_confusion = train_one_epoch(model, optimizer, log_interval)
    loss_history.extend(loss)
    train_acc_history.append(train_accuracy)
    
    # validation
    val_accuracy, val_confusion = check_val_accuracy(model)
    val_acc_history.append(val_accuracy)
    
    # learning rate schedule
    scheduler.step()
    
    print()
    print(f'* Train accuracy / confusion: {train_accuracy:.2f}% / {train_confusion.tolist()}, ')
    print(f'* Val   accuracy / confusion: {val_accuracy:.2f}% / {val_confusion.tolist()}')
    print()
    
# test
test_accuracy, test_confusion, test_debug = check_test_accuracy(model, repeat=30)

print(f'{"*"*40} Training Ends {"*"*40}')
print(f'- Test accuracy: {test_accuracy:.2f}%')
print()
print('- Confusion matrix:\n', test_confusion)
print()
print('- Debug table:')
pprint.pp(test_debug, indent=4)

# draw the training loss plot
draw_loss_plot(loss_history)
draw_accuracy_history(train_acc_history, val_acc_history)

# draw the confusion matrix
draw_confusion(test_confusion)

**************************************** Training Starts ****************************************

------------------------------ Epoch 001 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.095770
- Iter 024 / 025, Loss: 0.976993

* Train accuracy / confusion: 42.88% / [[297, 39, 17], [214, 39, 18], [150, 19, 7]], 
* Val   accuracy / confusion: 43.27% / [[42, 4, 0], [30, 3, 2], [20, 3, 0]]

------------------------------ Epoch 002 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.031583
- Iter 024 / 025, Loss: 1.054672

* Train accuracy / confusion: 45.75% / [[316, 34, 7], [213, 39, 12], [151, 17, 11]], 
* Val   accuracy / confusion: 44.23% / [[43, 3, 0], [30, 2, 3], [18, 4, 1]]

------------------------------ Epoch 003 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.066782
- Iter 024 / 025, Loss: 0.911509

* Train accuracy / confusion: 45.50% / [[310, 38, 12], [223, 42, 3], [138, 22, 12]], 
* Val   accuracy / confusion: 44.23% / [[41, 5, 0], [29, 5, 1], [21, 2, 0]]

------------------------------ Epoch 004 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.002660
- Iter 024 / 025, Loss: 1.121735

* Train accuracy / confusion: 45.38% / [[293, 40, 17], [207, 54, 11], [135, 27, 16]], 
* Val   accuracy / confusion: 45.19% / [[38, 7, 1], [25, 9, 1], [21, 2, 0]]

------------------------------ Epoch 005 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.058979
- Iter 024 / 025, Loss: 1.017595

* Train accuracy / confusion: 45.62% / [[293, 58, 7], [195, 62, 7], [128, 40, 10]], 
* Val   accuracy / confusion: 46.15% / [[36, 7, 3], [19, 11, 5], [18, 4, 1]]

------------------------------ Epoch 006 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.034961
- Iter 024 / 025, Loss: 1.043857

* Train accuracy / confusion: 48.38% / [[288, 52, 17], [167, 90, 11], [113, 53, 9]], 
* Val   accuracy / confusion: 49.04% / [[37, 8, 1], [21, 14, 0], [13, 10, 0]]

------------------------------ Epoch 007 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.937380
- Iter 024 / 025, Loss: 1.138952

* Train accuracy / confusion: 51.00% / [[291, 56, 14], [157, 93, 16], [102, 47, 24]], 
* Val   accuracy / confusion: 49.04% / [[34, 12, 0], [20, 15, 0], [9, 12, 2]]

------------------------------ Epoch 008 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.938843
- Iter 024 / 025, Loss: 1.097286

* Train accuracy / confusion: 52.75% / [[282, 63, 10], [142, 114, 13], [103, 47, 26]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [10, 21, 4], [9, 12, 2]]

------------------------------ Epoch 009 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.031000
- Iter 024 / 025, Loss: 0.930799

* Train accuracy / confusion: 52.88% / [[265, 78, 13], [121, 129, 23], [80, 62, 29]], 
* Val   accuracy / confusion: 48.08% / [[32, 14, 0], [16, 17, 2], [9, 13, 1]]

------------------------------ Epoch 010 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.041083
- Iter 024 / 025, Loss: 0.983541

* Train accuracy / confusion: 51.25% / [[261, 83, 9], [138, 113, 20], [80, 60, 36]], 
* Val   accuracy / confusion: 50.00% / [[33, 12, 1], [14, 14, 7], [11, 7, 5]]

------------------------------ Epoch 011 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.801915
- Iter 024 / 025, Loss: 0.902068

* Train accuracy / confusion: 55.88% / [[273, 68, 17], [116, 130, 24], [75, 53, 44]], 
* Val   accuracy / confusion: 48.08% / [[40, 5, 1], [23, 10, 2], [16, 7, 0]]

------------------------------ Epoch 012 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.888717
- Iter 024 / 025, Loss: 0.946282

* Train accuracy / confusion: 54.25% / [[280, 62, 14], [120, 114, 36], [72, 62, 40]], 
* Val   accuracy / confusion: 52.88% / [[32, 13, 1], [10, 22, 3], [11, 11, 1]]

------------------------------ Epoch 013 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.924545
- Iter 024 / 025, Loss: 0.889676

* Train accuracy / confusion: 48.88% / [[253, 87, 13], [135, 105, 31], [73, 70, 33]], 
* Val   accuracy / confusion: 47.12% / [[35, 9, 2], [18, 11, 6], [10, 10, 3]]

------------------------------ Epoch 014 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.935814
- Iter 024 / 025, Loss: 1.017739

* Train accuracy / confusion: 55.25% / [[263, 77, 12], [116, 132, 21], [68, 64, 47]], 
* Val   accuracy / confusion: 50.00% / [[34, 12, 0], [18, 15, 2], [11, 9, 3]]

------------------------------ Epoch 015 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.918739
- Iter 024 / 025, Loss: 0.901335

* Train accuracy / confusion: 54.00% / [[262, 67, 21], [127, 120, 27], [62, 64, 50]], 
* Val   accuracy / confusion: 52.88% / [[32, 10, 4], [14, 20, 1], [6, 14, 3]]

------------------------------ Epoch 016 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.914306
- Iter 024 / 025, Loss: 1.046786

* Train accuracy / confusion: 55.38% / [[259, 77, 21], [104, 130, 35], [58, 62, 54]], 
* Val   accuracy / confusion: 47.12% / [[31, 7, 8], [13, 11, 11], [5, 11, 7]]

------------------------------ Epoch 017 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.023721
- Iter 024 / 025, Loss: 0.927808

* Train accuracy / confusion: 53.75% / [[256, 79, 18], [110, 127, 34], [60, 69, 47]], 
* Val   accuracy / confusion: 52.88% / [[36, 7, 3], [16, 16, 3], [11, 9, 3]]

------------------------------ Epoch 018 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.840767
- Iter 024 / 025, Loss: 0.953867

* Train accuracy / confusion: 55.38% / [[268, 77, 15], [104, 127, 35], [60, 66, 48]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [10, 17, 8], [8, 10, 5]]

------------------------------ Epoch 019 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.893013
- Iter 024 / 025, Loss: 0.868188

* Train accuracy / confusion: 52.88% / [[251, 80, 24], [112, 124, 33], [56, 72, 48]], 
* Val   accuracy / confusion: 51.92% / [[34, 11, 1], [13, 19, 3], [9, 13, 1]]

------------------------------ Epoch 020 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.108368
- Iter 024 / 025, Loss: 0.832752

* Train accuracy / confusion: 56.75% / [[264, 76, 16], [107, 127, 35], [54, 58, 63]], 
* Val   accuracy / confusion: 50.00% / [[34, 8, 4], [17, 14, 4], [12, 7, 4]]

------------------------------ Epoch 021 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.951278
- Iter 024 / 025, Loss: 0.876543

* Train accuracy / confusion: 55.50% / [[262, 71, 19], [116, 113, 41], [64, 45, 69]], 
* Val   accuracy / confusion: 53.85% / [[32, 7, 7], [7, 18, 10], [6, 11, 6]]

------------------------------ Epoch 022 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.074057
- Iter 024 / 025, Loss: 0.913740

* Train accuracy / confusion: 54.50% / [[265, 80, 15], [106, 116, 44], [56, 63, 55]], 
* Val   accuracy / confusion: 52.88% / [[30, 10, 6], [9, 17, 9], [5, 10, 8]]

------------------------------ Epoch 023 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.960063
- Iter 024 / 025, Loss: 1.026090

* Train accuracy / confusion: 57.38% / [[265, 73, 17], [93, 136, 41], [48, 69, 58]], 
* Val   accuracy / confusion: 53.85% / [[30, 13, 3], [9, 20, 6], [6, 11, 6]]

------------------------------ Epoch 024 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.692984
- Iter 024 / 025, Loss: 0.904970

* Train accuracy / confusion: 57.38% / [[271, 71, 17], [105, 123, 38], [48, 62, 65]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [10, 20, 5], [5, 13, 5]]

------------------------------ Epoch 025 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.969665
- Iter 024 / 025, Loss: 0.791665

* Train accuracy / confusion: 58.12% / [[259, 75, 22], [95, 136, 34], [45, 64, 70]], 
* Val   accuracy / confusion: 57.69% / [[36, 8, 2], [14, 15, 6], [5, 9, 9]]

------------------------------ Epoch 026 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.907857
- Iter 024 / 025, Loss: 0.949617

* Train accuracy / confusion: 57.38% / [[258, 69, 25], [88, 135, 45], [52, 62, 66]], 
* Val   accuracy / confusion: 54.81% / [[30, 11, 5], [7, 18, 10], [4, 10, 9]]

------------------------------ Epoch 027 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.803160
- Iter 024 / 025, Loss: 0.912322

* Train accuracy / confusion: 59.88% / [[267, 74, 15], [94, 129, 45], [44, 49, 83]], 
* Val   accuracy / confusion: 60.58% / [[29, 11, 6], [2, 26, 7], [6, 9, 8]]

------------------------------ Epoch 028 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.918113
- Iter 024 / 025, Loss: 0.948703

* Train accuracy / confusion: 60.12% / [[266, 73, 18], [99, 137, 34], [45, 50, 78]], 
* Val   accuracy / confusion: 58.65% / [[33, 4, 9], [13, 15, 7], [8, 2, 13]]

------------------------------ Epoch 029 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.822462
- Iter 024 / 025, Loss: 0.968271

* Train accuracy / confusion: 58.88% / [[268, 64, 27], [91, 133, 44], [52, 51, 70]], 
* Val   accuracy / confusion: 57.69% / [[36, 7, 3], [12, 18, 5], [10, 7, 6]]

------------------------------ Epoch 030 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.906936
- Iter 024 / 025, Loss: 0.868895

* Train accuracy / confusion: 61.25% / [[286, 54, 17], [90, 140, 35], [50, 64, 64]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [9, 23, 3], [8, 11, 4]]

------------------------------ Epoch 031 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.822210
- Iter 024 / 025, Loss: 0.860826

* Train accuracy / confusion: 59.00% / [[271, 64, 20], [112, 122, 34], [50, 48, 79]], 
* Val   accuracy / confusion: 51.92% / [[27, 14, 5], [4, 18, 13], [5, 9, 9]]

------------------------------ Epoch 032 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.832338
- Iter 024 / 025, Loss: 0.876896

* Train accuracy / confusion: 57.62% / [[258, 69, 27], [92, 135, 42], [46, 63, 68]], 
* Val   accuracy / confusion: 51.92% / [[28, 12, 6], [6, 17, 12], [3, 11, 9]]

------------------------------ Epoch 033 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.898807
- Iter 024 / 025, Loss: 0.839457

* Train accuracy / confusion: 58.12% / [[258, 73, 27], [88, 136, 42], [48, 57, 71]], 
* Val   accuracy / confusion: 52.88% / [[28, 15, 3], [7, 20, 8], [7, 9, 7]]

------------------------------ Epoch 034 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.822986
- Iter 024 / 025, Loss: 0.979232

* Train accuracy / confusion: 60.00% / [[258, 79, 15], [95, 141, 35], [43, 53, 81]], 
* Val   accuracy / confusion: 58.65% / [[32, 8, 6], [7, 19, 9], [6, 7, 10]]

------------------------------ Epoch 035 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.774629
- Iter 024 / 025, Loss: 0.825390

* Train accuracy / confusion: 60.00% / [[255, 82, 22], [86, 137, 43], [39, 48, 88]], 
* Val   accuracy / confusion: 54.81% / [[30, 10, 6], [9, 18, 8], [7, 7, 9]]

------------------------------ Epoch 036 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.691596
- Iter 024 / 025, Loss: 0.990316

* Train accuracy / confusion: 61.62% / [[271, 68, 14], [89, 143, 38], [41, 57, 79]], 
* Val   accuracy / confusion: 46.15% / [[33, 8, 5], [17, 10, 8], [5, 13, 5]]

------------------------------ Epoch 037 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.917310
- Iter 024 / 025, Loss: 0.852908

* Train accuracy / confusion: 61.88% / [[273, 58, 28], [91, 142, 38], [41, 49, 80]], 
* Val   accuracy / confusion: 51.92% / [[30, 12, 4], [9, 18, 8], [5, 12, 6]]

------------------------------ Epoch 038 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.882526
- Iter 024 / 025, Loss: 1.004986

* Train accuracy / confusion: 62.00% / [[273, 61, 21], [92, 141, 39], [42, 49, 82]], 
* Val   accuracy / confusion: 55.77% / [[31, 9, 6], [7, 17, 11], [3, 10, 10]]

------------------------------ Epoch 039 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.740295
- Iter 024 / 025, Loss: 0.686148

* Train accuracy / confusion: 60.12% / [[278, 62, 19], [95, 128, 41], [45, 57, 75]], 
* Val   accuracy / confusion: 49.04% / [[27, 13, 6], [11, 21, 3], [4, 16, 3]]

------------------------------ Epoch 040 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.903119
- Iter 024 / 025, Loss: 0.757865

* Train accuracy / confusion: 61.38% / [[272, 69, 15], [85, 142, 40], [43, 57, 77]], 
* Val   accuracy / confusion: 57.69% / [[36, 7, 3], [12, 20, 3], [8, 11, 4]]

------------------------------ Epoch 041 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.865541
- Iter 024 / 025, Loss: 0.806718

* Train accuracy / confusion: 60.75% / [[259, 75, 20], [84, 146, 35], [39, 61, 81]], 
* Val   accuracy / confusion: 55.77% / [[37, 4, 5], [11, 16, 8], [7, 11, 5]]

------------------------------ Epoch 042 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.686485
- Iter 024 / 025, Loss: 0.764748

* Train accuracy / confusion: 61.75% / [[265, 69, 23], [77, 146, 47], [45, 45, 83]], 
* Val   accuracy / confusion: 50.96% / [[31, 8, 7], [8, 18, 9], [7, 12, 4]]

------------------------------ Epoch 043 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.871630
- Iter 024 / 025, Loss: 0.787186

* Train accuracy / confusion: 60.00% / [[269, 67, 22], [99, 133, 37], [46, 49, 78]], 
* Val   accuracy / confusion: 51.92% / [[34, 7, 5], [10, 14, 11], [4, 13, 6]]

------------------------------ Epoch 044 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.707618
- Iter 024 / 025, Loss: 0.927612

* Train accuracy / confusion: 60.88% / [[281, 63, 16], [99, 122, 48], [47, 40, 84]], 
* Val   accuracy / confusion: 58.65% / [[31, 12, 3], [7, 23, 5], [7, 9, 7]]

------------------------------ Epoch 045 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.878324
- Iter 024 / 025, Loss: 0.935147

* Train accuracy / confusion: 61.25% / [[277, 64, 9], [90, 141, 40], [54, 53, 72]], 
* Val   accuracy / confusion: 53.85% / [[39, 5, 2], [16, 14, 5], [14, 6, 3]]

------------------------------ Epoch 046 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.654498
- Iter 024 / 025, Loss: 1.017362

* Train accuracy / confusion: 64.50% / [[285, 60, 13], [84, 152, 34], [42, 51, 79]], 
* Val   accuracy / confusion: 52.88% / [[32, 12, 2], [8, 19, 8], [7, 12, 4]]

------------------------------ Epoch 047 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.741003
- Iter 024 / 025, Loss: 0.851581

* Train accuracy / confusion: 62.62% / [[284, 57, 13], [94, 145, 29], [51, 55, 72]], 
* Val   accuracy / confusion: 53.85% / [[30, 12, 4], [9, 19, 7], [6, 10, 7]]

------------------------------ Epoch 048 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.703737
- Iter 024 / 025, Loss: 0.846688

* Train accuracy / confusion: 63.75% / [[281, 65, 16], [85, 137, 44], [32, 48, 92]], 
* Val   accuracy / confusion: 50.00% / [[27, 9, 10], [7, 14, 14], [4, 8, 11]]

------------------------------ Epoch 049 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.854030
- Iter 024 / 025, Loss: 0.748246

* Train accuracy / confusion: 62.62% / [[268, 63, 22], [89, 145, 37], [40, 48, 88]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [15, 13, 7], [9, 6, 8]]

------------------------------ Epoch 050 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.759046
- Iter 024 / 025, Loss: 1.008546

* Train accuracy / confusion: 61.38% / [[281, 57, 21], [93, 128, 45], [42, 51, 82]], 
* Val   accuracy / confusion: 57.69% / [[32, 8, 6], [7, 20, 8], [4, 11, 8]]

------------------------------ Epoch 051 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.799529
- Iter 024 / 025, Loss: 0.750256

* Train accuracy / confusion: 61.75% / [[270, 67, 18], [87, 144, 39], [44, 51, 80]], 
* Val   accuracy / confusion: 51.92% / [[30, 13, 3], [7, 21, 7], [6, 14, 3]]

------------------------------ Epoch 052 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.829067
- Iter 024 / 025, Loss: 0.702338

* Train accuracy / confusion: 65.88% / [[285, 60, 14], [79, 151, 37], [37, 46, 91]], 
* Val   accuracy / confusion: 53.85% / [[36, 7, 3], [13, 19, 3], [11, 11, 1]]

------------------------------ Epoch 053 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.768227
- Iter 024 / 025, Loss: 0.675755

* Train accuracy / confusion: 65.50% / [[285, 60, 12], [85, 150, 34], [34, 51, 89]], 
* Val   accuracy / confusion: 49.04% / [[32, 11, 3], [13, 16, 6], [9, 11, 3]]

------------------------------ Epoch 054 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.736292
- Iter 024 / 025, Loss: 0.834154

* Train accuracy / confusion: 65.75% / [[280, 57, 18], [75, 155, 38], [41, 45, 91]], 
* Val   accuracy / confusion: 51.92% / [[32, 10, 4], [11, 15, 9], [10, 6, 7]]

------------------------------ Epoch 055 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.832979
- Iter 024 / 025, Loss: 0.721683

* Train accuracy / confusion: 65.00% / [[278, 62, 19], [82, 144, 41], [37, 39, 98]], 
* Val   accuracy / confusion: 51.92% / [[33, 10, 3], [11, 14, 10], [6, 10, 7]]

------------------------------ Epoch 056 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.697636
- Iter 024 / 025, Loss: 0.933554

* Train accuracy / confusion: 64.50% / [[287, 50, 19], [76, 148, 44], [50, 45, 81]], 
* Val   accuracy / confusion: 46.15% / [[34, 7, 5], [16, 11, 8], [10, 10, 3]]

------------------------------ Epoch 057 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.865665
- Iter 024 / 025, Loss: 0.713121

* Train accuracy / confusion: 64.62% / [[287, 50, 17], [89, 143, 35], [44, 48, 87]], 
* Val   accuracy / confusion: 56.73% / [[34, 7, 5], [13, 19, 3], [5, 12, 6]]

------------------------------ Epoch 058 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.715476
- Iter 024 / 025, Loss: 0.674321

* Train accuracy / confusion: 69.25% / [[289, 53, 17], [64, 164, 39], [36, 37, 101]], 
* Val   accuracy / confusion: 58.65% / [[34, 7, 5], [12, 18, 5], [8, 6, 9]]

------------------------------ Epoch 059 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.877957
- Iter 024 / 025, Loss: 1.043611

* Train accuracy / confusion: 63.25% / [[274, 61, 19], [86, 147, 36], [40, 52, 85]], 
* Val   accuracy / confusion: 54.81% / [[32, 7, 7], [12, 13, 10], [4, 7, 12]]

------------------------------ Epoch 060 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.930636
- Iter 024 / 025, Loss: 0.877667

* Train accuracy / confusion: 66.75% / [[299, 40, 17], [95, 139, 34], [34, 46, 96]], 
* Val   accuracy / confusion: 53.85% / [[35, 8, 3], [12, 14, 9], [6, 10, 7]]

------------------------------ Epoch 061 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.803548
- Iter 024 / 025, Loss: 0.883447

* Train accuracy / confusion: 62.75% / [[275, 65, 15], [84, 138, 48], [38, 48, 89]], 
* Val   accuracy / confusion: 57.69% / [[32, 10, 4], [7, 20, 8], [3, 12, 8]]

------------------------------ Epoch 062 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.727328
- Iter 024 / 025, Loss: 0.729635

* Train accuracy / confusion: 66.00% / [[286, 55, 18], [86, 146, 35], [34, 44, 96]], 
* Val   accuracy / confusion: 56.73% / [[33, 7, 6], [7, 17, 11], [8, 6, 9]]

------------------------------ Epoch 063 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.782396
- Iter 024 / 025, Loss: 0.882481

* Train accuracy / confusion: 66.25% / [[282, 51, 23], [78, 153, 37], [38, 43, 95]], 
* Val   accuracy / confusion: 51.92% / [[37, 4, 5], [13, 11, 11], [7, 10, 6]]

------------------------------ Epoch 064 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.905705
- Iter 024 / 025, Loss: 0.703360

* Train accuracy / confusion: 65.88% / [[288, 49, 18], [76, 162, 32], [37, 61, 77]], 
* Val   accuracy / confusion: 55.77% / [[33, 9, 4], [12, 19, 4], [7, 10, 6]]

------------------------------ Epoch 065 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.899202
- Iter 024 / 025, Loss: 0.757593

* Train accuracy / confusion: 66.75% / [[286, 51, 19], [79, 156, 31], [38, 48, 92]], 
* Val   accuracy / confusion: 47.12% / [[34, 10, 2], [18, 13, 4], [13, 8, 2]]

------------------------------ Epoch 066 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.733518
- Iter 024 / 025, Loss: 0.736036

* Train accuracy / confusion: 66.25% / [[286, 53, 16], [82, 155, 33], [38, 48, 89]], 
* Val   accuracy / confusion: 50.00% / [[30, 10, 6], [7, 15, 13], [4, 12, 7]]

------------------------------ Epoch 067 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.597785
- Iter 024 / 025, Loss: 0.733558

* Train accuracy / confusion: 66.12% / [[287, 54, 16], [87, 148, 33], [39, 42, 94]], 
* Val   accuracy / confusion: 50.96% / [[30, 12, 4], [13, 16, 6], [2, 14, 7]]

------------------------------ Epoch 068 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.669500
- Iter 024 / 025, Loss: 0.684304

* Train accuracy / confusion: 66.62% / [[284, 53, 18], [72, 154, 44], [36, 44, 95]], 
* Val   accuracy / confusion: 49.04% / [[28, 11, 7], [10, 15, 10], [3, 12, 8]]

------------------------------ Epoch 069 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.753471
- Iter 024 / 025, Loss: 0.707353

* Train accuracy / confusion: 66.25% / [[282, 59, 19], [79, 147, 41], [34, 38, 101]], 
* Val   accuracy / confusion: 49.04% / [[35, 6, 5], [15, 10, 10], [12, 5, 6]]

------------------------------ Epoch 070 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.589282
- Iter 024 / 025, Loss: 0.564541

* Train accuracy / confusion: 67.25% / [[292, 39, 24], [77, 152, 42], [35, 45, 94]], 
* Val   accuracy / confusion: 51.92% / [[33, 7, 6], [13, 13, 9], [7, 8, 8]]

------------------------------ Epoch 071 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.609082
- Iter 024 / 025, Loss: 0.707925

* Train accuracy / confusion: 67.50% / [[287, 50, 17], [76, 155, 39], [39, 39, 98]], 
* Val   accuracy / confusion: 53.85% / [[33, 10, 3], [9, 17, 9], [8, 9, 6]]

------------------------------ Epoch 072 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.740941
- Iter 024 / 025, Loss: 0.894174

* Train accuracy / confusion: 68.12% / [[286, 56, 18], [63, 163, 36], [36, 46, 96]], 
* Val   accuracy / confusion: 52.88% / [[29, 10, 7], [7, 17, 11], [4, 10, 9]]

------------------------------ Epoch 073 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.900526
- Iter 024 / 025, Loss: 0.586693

* Train accuracy / confusion: 66.00% / [[276, 58, 25], [79, 162, 27], [37, 46, 90]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [10, 17, 8], [5, 11, 7]]

------------------------------ Epoch 074 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.661704
- Iter 024 / 025, Loss: 0.750860

* Train accuracy / confusion: 71.38% / [[293, 52, 12], [66, 173, 30], [31, 38, 105]], 
* Val   accuracy / confusion: 51.92% / [[30, 13, 3], [12, 16, 7], [4, 11, 8]]

------------------------------ Epoch 075 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.839086
- Iter 024 / 025, Loss: 0.731022

* Train accuracy / confusion: 67.50% / [[290, 50, 15], [75, 143, 49], [36, 35, 107]], 
* Val   accuracy / confusion: 54.81% / [[34, 9, 3], [14, 17, 4], [7, 10, 6]]

------------------------------ Epoch 076 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.670980
- Iter 024 / 025, Loss: 0.706297

* Train accuracy / confusion: 70.75% / [[296, 50, 12], [58, 167, 39], [37, 38, 103]], 
* Val   accuracy / confusion: 55.77% / [[28, 12, 6], [5, 23, 7], [4, 12, 7]]

------------------------------ Epoch 077 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.812978
- Iter 024 / 025, Loss: 0.723982

* Train accuracy / confusion: 66.12% / [[280, 55, 23], [72, 159, 39], [34, 48, 90]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [7, 23, 5], [10, 9, 4]]

------------------------------ Epoch 078 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.624410
- Iter 024 / 025, Loss: 0.832792

* Train accuracy / confusion: 67.25% / [[282, 58, 11], [83, 150, 39], [32, 39, 106]], 
* Val   accuracy / confusion: 57.69% / [[34, 6, 6], [11, 17, 7], [7, 7, 9]]

------------------------------ Epoch 079 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.781331
- Iter 024 / 025, Loss: 0.830410

* Train accuracy / confusion: 66.75% / [[296, 47, 15], [74, 150, 41], [36, 53, 88]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [12, 15, 8], [2, 11, 10]]

------------------------------ Epoch 080 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.545012
- Iter 024 / 025, Loss: 0.699631

* Train accuracy / confusion: 68.00% / [[293, 44, 13], [80, 159, 31], [32, 56, 92]], 
* Val   accuracy / confusion: 51.92% / [[28, 10, 8], [9, 17, 9], [5, 9, 9]]

------------------------------ Epoch 081 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.467258
- Iter 024 / 025, Loss: 0.646486

* Train accuracy / confusion: 69.38% / [[291, 45, 16], [65, 161, 46], [28, 45, 103]], 
* Val   accuracy / confusion: 50.96% / [[33, 6, 7], [9, 11, 15], [7, 7, 9]]

------------------------------ Epoch 082 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.619733
- Iter 024 / 025, Loss: 0.642232

* Train accuracy / confusion: 71.50% / [[295, 46, 13], [72, 168, 31], [33, 33, 109]], 
* Val   accuracy / confusion: 48.08% / [[28, 13, 5], [10, 19, 6], [7, 13, 3]]

------------------------------ Epoch 083 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.771212
- Iter 024 / 025, Loss: 0.790526

* Train accuracy / confusion: 72.62% / [[305, 40, 12], [67, 171, 31], [26, 43, 105]], 
* Val   accuracy / confusion: 54.81% / [[30, 10, 6], [6, 19, 10], [4, 11, 8]]

------------------------------ Epoch 084 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.482101
- Iter 024 / 025, Loss: 0.659828

* Train accuracy / confusion: 69.25% / [[284, 54, 18], [61, 171, 35], [37, 41, 99]], 
* Val   accuracy / confusion: 50.96% / [[31, 8, 7], [11, 15, 9], [5, 11, 7]]

------------------------------ Epoch 085 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 1.046821
- Iter 024 / 025, Loss: 0.579974

* Train accuracy / confusion: 70.12% / [[291, 49, 16], [73, 167, 24], [35, 42, 103]], 
* Val   accuracy / confusion: 57.69% / [[35, 6, 5], [7, 20, 8], [13, 5, 5]]

------------------------------ Epoch 086 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.709248
- Iter 024 / 025, Loss: 0.784021

* Train accuracy / confusion: 72.25% / [[292, 46, 15], [64, 176, 30], [31, 36, 110]], 
* Val   accuracy / confusion: 53.85% / [[33, 8, 5], [13, 14, 8], [9, 5, 9]]

------------------------------ Epoch 087 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.821578
- Iter 024 / 025, Loss: 0.820356

* Train accuracy / confusion: 67.62% / [[286, 52, 17], [71, 155, 41], [37, 41, 100]], 
* Val   accuracy / confusion: 57.69% / [[30, 11, 5], [8, 21, 6], [6, 8, 9]]

------------------------------ Epoch 088 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.639146
- Iter 024 / 025, Loss: 0.703937

* Train accuracy / confusion: 70.88% / [[296, 43, 14], [77, 162, 31], [31, 37, 109]], 
* Val   accuracy / confusion: 57.69% / [[29, 11, 6], [5, 16, 14], [2, 6, 15]]

------------------------------ Epoch 089 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.618960
- Iter 024 / 025, Loss: 0.652898

* Train accuracy / confusion: 71.25% / [[295, 47, 16], [68, 165, 33], [29, 37, 110]], 
* Val   accuracy / confusion: 57.69% / [[31, 13, 2], [7, 24, 4], [6, 12, 5]]

------------------------------ Epoch 090 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.626933
- Iter 024 / 025, Loss: 0.778511

* Train accuracy / confusion: 71.00% / [[289, 54, 16], [67, 168, 34], [23, 38, 111]], 
* Val   accuracy / confusion: 52.88% / [[31, 11, 4], [12, 18, 5], [3, 14, 6]]

------------------------------ Epoch 091 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.673781
- Iter 024 / 025, Loss: 0.958539

* Train accuracy / confusion: 70.12% / [[290, 48, 14], [73, 167, 30], [33, 41, 104]], 
* Val   accuracy / confusion: 50.96% / [[35, 6, 5], [13, 15, 7], [11, 9, 3]]

------------------------------ Epoch 092 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.595503
- Iter 024 / 025, Loss: 0.559473

* Train accuracy / confusion: 72.25% / [[291, 48, 15], [64, 177, 30], [32, 33, 110]], 
* Val   accuracy / confusion: 59.62% / [[32, 11, 3], [8, 21, 6], [6, 8, 9]]

------------------------------ Epoch 093 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.602722
- Iter 024 / 025, Loss: 0.506540

* Train accuracy / confusion: 71.12% / [[301, 54, 5], [66, 164, 34], [29, 43, 104]], 
* Val   accuracy / confusion: 50.96% / [[31, 10, 5], [10, 19, 6], [11, 9, 3]]

------------------------------ Epoch 094 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.517909
- Iter 024 / 025, Loss: 0.567057

* Train accuracy / confusion: 72.38% / [[297, 44, 14], [62, 174, 36], [26, 39, 108]], 
* Val   accuracy / confusion: 50.00% / [[32, 12, 2], [11, 16, 8], [8, 11, 4]]

------------------------------ Epoch 095 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.748147
- Iter 024 / 025, Loss: 0.622079

* Train accuracy / confusion: 70.88% / [[293, 41, 19], [61, 166, 44], [29, 39, 108]], 
* Val   accuracy / confusion: 55.77% / [[30, 7, 9], [7, 18, 10], [5, 8, 10]]

------------------------------ Epoch 096 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.650229
- Iter 024 / 025, Loss: 0.946290

* Train accuracy / confusion: 71.00% / [[290, 51, 15], [54, 173, 42], [27, 43, 105]], 
* Val   accuracy / confusion: 60.58% / [[34, 8, 4], [11, 18, 6], [5, 7, 11]]

------------------------------ Epoch 097 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.710575
- Iter 024 / 025, Loss: 0.854271

* Train accuracy / confusion: 72.38% / [[297, 47, 13], [67, 166, 36], [23, 35, 116]], 
* Val   accuracy / confusion: 57.69% / [[36, 5, 5], [10, 18, 7], [10, 7, 6]]

------------------------------ Epoch 098 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.770572
- Iter 024 / 025, Loss: 0.621287

* Train accuracy / confusion: 72.62% / [[300, 39, 13], [74, 164, 35], [19, 39, 117]], 
* Val   accuracy / confusion: 49.04% / [[31, 8, 7], [8, 12, 15], [7, 8, 8]]

------------------------------ Epoch 099 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.711532
- Iter 024 / 025, Loss: 0.496686

* Train accuracy / confusion: 70.88% / [[295, 44, 17], [64, 167, 37], [34, 37, 105]], 
* Val   accuracy / confusion: 58.65% / [[33, 10, 3], [10, 21, 4], [5, 11, 7]]

------------------------------ Epoch 100 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.765013
- Iter 024 / 025, Loss: 0.604593

* Train accuracy / confusion: 70.75% / [[293, 45, 19], [64, 169, 35], [33, 38, 104]], 
* Val   accuracy / confusion: 51.92% / [[27, 14, 5], [8, 22, 5], [8, 10, 5]]

------------------------------ Epoch 101 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.699413
- Iter 024 / 025, Loss: 0.703809

* Train accuracy / confusion: 72.38% / [[295, 49, 11], [62, 173, 34], [23, 42, 111]], 
* Val   accuracy / confusion: 52.88% / [[30, 9, 7], [8, 20, 7], [6, 12, 5]]

------------------------------ Epoch 102 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.518308
- Iter 024 / 025, Loss: 0.710436

* Train accuracy / confusion: 71.12% / [[293, 40, 19], [56, 172, 43], [24, 49, 104]], 
* Val   accuracy / confusion: 51.92% / [[34, 9, 3], [10, 15, 10], [10, 8, 5]]

------------------------------ Epoch 103 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.518707
- Iter 024 / 025, Loss: 0.527888

* Train accuracy / confusion: 74.12% / [[302, 42, 11], [69, 178, 21], [28, 36, 113]], 
* Val   accuracy / confusion: 55.77% / [[30, 10, 6], [8, 19, 8], [4, 10, 9]]

------------------------------ Epoch 104 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.739240
- Iter 024 / 025, Loss: 0.762757

* Train accuracy / confusion: 73.00% / [[298, 42, 13], [56, 179, 37], [21, 47, 107]], 
* Val   accuracy / confusion: 58.65% / [[28, 11, 7], [5, 20, 10], [3, 7, 13]]

------------------------------ Epoch 105 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.694473
- Iter 024 / 025, Loss: 0.515828

* Train accuracy / confusion: 72.62% / [[302, 40, 18], [64, 169, 35], [29, 33, 110]], 
* Val   accuracy / confusion: 53.85% / [[36, 4, 6], [15, 10, 10], [7, 6, 10]]

------------------------------ Epoch 106 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.490670
- Iter 024 / 025, Loss: 0.568535

* Train accuracy / confusion: 71.25% / [[301, 41, 13], [64, 166, 38], [31, 43, 103]], 
* Val   accuracy / confusion: 56.73% / [[28, 14, 4], [9, 23, 3], [7, 8, 8]]

------------------------------ Epoch 107 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.563825
- Iter 024 / 025, Loss: 0.671495

* Train accuracy / confusion: 72.38% / [[288, 48, 18], [59, 176, 35], [21, 40, 115]], 
* Val   accuracy / confusion: 55.77% / [[33, 8, 5], [7, 17, 11], [8, 7, 8]]

------------------------------ Epoch 108 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.685663
- Iter 024 / 025, Loss: 0.513165

* Train accuracy / confusion: 75.38% / [[303, 41, 10], [60, 178, 30], [25, 31, 122]], 
* Val   accuracy / confusion: 50.96% / [[28, 12, 6], [7, 16, 12], [5, 9, 9]]

------------------------------ Epoch 109 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.819958
- Iter 024 / 025, Loss: 0.724980

* Train accuracy / confusion: 72.25% / [[297, 47, 14], [65, 172, 32], [33, 31, 109]], 
* Val   accuracy / confusion: 59.62% / [[34, 10, 2], [10, 22, 3], [6, 11, 6]]

------------------------------ Epoch 110 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.661283
- Iter 024 / 025, Loss: 0.567985

* Train accuracy / confusion: 72.38% / [[285, 56, 12], [61, 182, 26], [25, 41, 112]], 
* Val   accuracy / confusion: 57.69% / [[29, 13, 4], [4, 25, 6], [5, 12, 6]]

------------------------------ Epoch 111 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.542509
- Iter 024 / 025, Loss: 0.715369

* Train accuracy / confusion: 73.12% / [[308, 36, 13], [60, 169, 40], [22, 44, 108]], 
* Val   accuracy / confusion: 49.04% / [[29, 9, 8], [12, 12, 11], [8, 5, 10]]

------------------------------ Epoch 112 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.756250
- Iter 024 / 025, Loss: 0.595040

* Train accuracy / confusion: 75.12% / [[301, 42, 16], [52, 188, 28], [25, 36, 112]], 
* Val   accuracy / confusion: 53.85% / [[31, 8, 7], [10, 14, 11], [6, 6, 11]]

------------------------------ Epoch 113 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.622022
- Iter 024 / 025, Loss: 0.817157

* Train accuracy / confusion: 75.88% / [[307, 39, 11], [60, 184, 26], [27, 30, 116]], 
* Val   accuracy / confusion: 52.88% / [[32, 9, 5], [10, 18, 7], [8, 10, 5]]

------------------------------ Epoch 114 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.444443
- Iter 024 / 025, Loss: 0.809465

* Train accuracy / confusion: 73.50% / [[308, 34, 18], [60, 179, 31], [27, 42, 101]], 
* Val   accuracy / confusion: 54.81% / [[34, 9, 3], [10, 18, 7], [9, 9, 5]]

------------------------------ Epoch 115 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.623810
- Iter 024 / 025, Loss: 0.635122

* Train accuracy / confusion: 74.62% / [[306, 37, 14], [71, 168, 28], [27, 26, 123]], 
* Val   accuracy / confusion: 63.46% / [[34, 8, 4], [8, 20, 7], [7, 4, 12]]

------------------------------ Epoch 116 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.540337
- Iter 024 / 025, Loss: 0.665304

* Train accuracy / confusion: 75.88% / [[309, 38, 11], [61, 183, 27], [25, 31, 115]], 
* Val   accuracy / confusion: 52.88% / [[30, 8, 8], [6, 18, 11], [7, 9, 7]]

------------------------------ Epoch 117 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.455862
- Iter 024 / 025, Loss: 0.411326

* Train accuracy / confusion: 75.25% / [[295, 49, 10], [56, 183, 28], [24, 31, 124]], 
* Val   accuracy / confusion: 51.92% / [[28, 12, 6], [10, 17, 8], [3, 11, 9]]

------------------------------ Epoch 118 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.569587
- Iter 024 / 025, Loss: 0.483479

* Train accuracy / confusion: 75.50% / [[305, 39, 11], [66, 178, 25], [19, 36, 121]], 
* Val   accuracy / confusion: 52.88% / [[31, 6, 9], [10, 16, 9], [4, 11, 8]]

------------------------------ Epoch 119 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.515224
- Iter 024 / 025, Loss: 0.500537

* Train accuracy / confusion: 76.50% / [[309, 35, 10], [62, 183, 23], [26, 32, 120]], 
* Val   accuracy / confusion: 51.92% / [[30, 12, 4], [11, 19, 5], [9, 9, 5]]

------------------------------ Epoch 120 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.685788
- Iter 024 / 025, Loss: 0.753677

* Train accuracy / confusion: 76.25% / [[313, 27, 18], [63, 173, 28], [25, 29, 124]], 
* Val   accuracy / confusion: 50.00% / [[30, 4, 12], [10, 9, 16], [7, 3, 13]]

------------------------------ Epoch 121 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.475642
- Iter 024 / 025, Loss: 0.650733

* Train accuracy / confusion: 78.12% / [[309, 35, 7], [60, 187, 24], [22, 27, 129]], 
* Val   accuracy / confusion: 51.92% / [[33, 8, 5], [16, 12, 7], [8, 6, 9]]

------------------------------ Epoch 122 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.688769
- Iter 024 / 025, Loss: 0.621288

* Train accuracy / confusion: 76.50% / [[296, 45, 12], [63, 180, 27], [17, 24, 136]], 
* Val   accuracy / confusion: 52.88% / [[34, 8, 4], [13, 14, 8], [7, 9, 7]]

------------------------------ Epoch 123 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.599223
- Iter 024 / 025, Loss: 0.530782

* Train accuracy / confusion: 78.62% / [[304, 38, 10], [47, 199, 26], [24, 26, 126]], 
* Val   accuracy / confusion: 50.96% / [[28, 13, 5], [9, 18, 8], [9, 7, 7]]

------------------------------ Epoch 124 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.564701
- Iter 024 / 025, Loss: 0.742983

* Train accuracy / confusion: 76.88% / [[297, 43, 15], [47, 198, 26], [20, 34, 120]], 
* Val   accuracy / confusion: 55.77% / [[33, 4, 9], [9, 14, 12], [8, 4, 11]]

------------------------------ Epoch 125 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.470300
- Iter 024 / 025, Loss: 0.565653

* Train accuracy / confusion: 76.62% / [[312, 32, 14], [60, 175, 34], [21, 26, 126]], 
* Val   accuracy / confusion: 60.58% / [[33, 10, 3], [7, 20, 8], [5, 8, 10]]

------------------------------ Epoch 126 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.720758
- Iter 024 / 025, Loss: 0.567899

* Train accuracy / confusion: 76.88% / [[311, 36, 12], [56, 184, 25], [27, 29, 120]], 
* Val   accuracy / confusion: 58.65% / [[30, 10, 6], [7, 20, 8], [7, 5, 11]]

------------------------------ Epoch 127 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.475809
- Iter 024 / 025, Loss: 0.732683

* Train accuracy / confusion: 77.75% / [[304, 40, 10], [51, 193, 25], [17, 35, 125]], 
* Val   accuracy / confusion: 57.69% / [[33, 9, 4], [13, 17, 5], [10, 3, 10]]

------------------------------ Epoch 128 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.766925
- Iter 024 / 025, Loss: 0.650235

* Train accuracy / confusion: 77.38% / [[306, 29, 14], [61, 181, 31], [17, 29, 132]], 
* Val   accuracy / confusion: 52.88% / [[29, 10, 7], [6, 11, 18], [4, 4, 15]]

------------------------------ Epoch 129 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.398827
- Iter 024 / 025, Loss: 0.640211

* Train accuracy / confusion: 77.12% / [[315, 31, 8], [56, 181, 32], [17, 39, 121]], 
* Val   accuracy / confusion: 56.73% / [[36, 6, 4], [10, 17, 8], [9, 8, 6]]

------------------------------ Epoch 130 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.651271
- Iter 024 / 025, Loss: 0.536663

* Train accuracy / confusion: 79.00% / [[307, 36, 13], [51, 193, 28], [15, 25, 132]], 
* Val   accuracy / confusion: 50.96% / [[33, 9, 4], [12, 15, 8], [11, 7, 5]]

------------------------------ Epoch 131 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.581400
- Iter 024 / 025, Loss: 0.404398

* Train accuracy / confusion: 76.62% / [[301, 45, 12], [50, 188, 29], [19, 32, 124]], 
* Val   accuracy / confusion: 56.73% / [[36, 7, 3], [13, 15, 7], [6, 9, 8]]

------------------------------ Epoch 132 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.409798
- Iter 024 / 025, Loss: 0.610211

* Train accuracy / confusion: 79.88% / [[309, 33, 13], [42, 197, 28], [20, 25, 133]], 
* Val   accuracy / confusion: 56.73% / [[29, 12, 5], [8, 23, 4], [5, 11, 7]]

------------------------------ Epoch 133 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.430534
- Iter 024 / 025, Loss: 0.545669

* Train accuracy / confusion: 79.62% / [[317, 27, 11], [42, 199, 28], [16, 39, 121]], 
* Val   accuracy / confusion: 57.69% / [[32, 12, 2], [11, 19, 5], [4, 10, 9]]

------------------------------ Epoch 134 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.591260
- Iter 024 / 025, Loss: 0.640533

* Train accuracy / confusion: 79.62% / [[313, 34, 12], [43, 191, 31], [20, 23, 133]], 
* Val   accuracy / confusion: 53.85% / [[28, 8, 10], [5, 18, 12], [4, 9, 10]]

------------------------------ Epoch 135 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.368290
- Iter 024 / 025, Loss: 0.553894

* Train accuracy / confusion: 79.00% / [[312, 23, 16], [47, 194, 31], [24, 27, 126]], 
* Val   accuracy / confusion: 63.46% / [[33, 8, 5], [7, 20, 8], [9, 1, 13]]

------------------------------ Epoch 136 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.710139
- Iter 024 / 025, Loss: 0.503905

* Train accuracy / confusion: 79.62% / [[311, 29, 14], [50, 189, 30], [19, 21, 137]], 
* Val   accuracy / confusion: 58.65% / [[29, 11, 6], [3, 23, 9], [8, 6, 9]]

------------------------------ Epoch 137 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.506732
- Iter 024 / 025, Loss: 0.625041

* Train accuracy / confusion: 77.25% / [[297, 47, 15], [50, 191, 25], [18, 27, 130]], 
* Val   accuracy / confusion: 50.00% / [[31, 9, 6], [10, 16, 9], [5, 13, 5]]

------------------------------ Epoch 138 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.577733
- Iter 024 / 025, Loss: 0.580279

* Train accuracy / confusion: 79.25% / [[310, 28, 18], [42, 197, 28], [22, 28, 127]], 
* Val   accuracy / confusion: 52.88% / [[28, 11, 7], [8, 18, 9], [6, 8, 9]]

------------------------------ Epoch 139 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.460518
- Iter 024 / 025, Loss: 0.496923

* Train accuracy / confusion: 80.25% / [[313, 39, 5], [41, 196, 33], [24, 16, 133]], 
* Val   accuracy / confusion: 53.85% / [[31, 12, 3], [14, 13, 8], [6, 5, 12]]

------------------------------ Epoch 140 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.544169
- Iter 024 / 025, Loss: 0.484620

* Train accuracy / confusion: 80.62% / [[317, 31, 5], [46, 197, 27], [16, 30, 131]], 
* Val   accuracy / confusion: 56.73% / [[32, 7, 7], [7, 21, 7], [10, 7, 6]]

------------------------------ Epoch 141 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.340261
- Iter 024 / 025, Loss: 0.706111

* Train accuracy / confusion: 79.25% / [[311, 36, 8], [49, 192, 27], [15, 31, 131]], 
* Val   accuracy / confusion: 55.77% / [[32, 7, 7], [9, 15, 11], [6, 6, 11]]

------------------------------ Epoch 142 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.625786
- Iter 024 / 025, Loss: 0.357516

* Train accuracy / confusion: 78.12% / [[317, 28, 13], [52, 184, 29], [21, 32, 124]], 
* Val   accuracy / confusion: 60.58% / [[33, 8, 5], [6, 22, 7], [7, 8, 8]]

------------------------------ Epoch 143 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.289407
- Iter 024 / 025, Loss: 0.424693

* Train accuracy / confusion: 80.75% / [[319, 27, 11], [37, 204, 26], [13, 40, 123]], 
* Val   accuracy / confusion: 57.69% / [[32, 6, 8], [6, 16, 13], [2, 9, 12]]

------------------------------ Epoch 144 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.554239
- Iter 024 / 025, Loss: 0.361180

* Train accuracy / confusion: 79.88% / [[318, 20, 16], [45, 192, 34], [17, 29, 129]], 
* Val   accuracy / confusion: 59.62% / [[30, 9, 7], [4, 21, 10], [4, 8, 11]]

------------------------------ Epoch 145 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.680426
- Iter 024 / 025, Loss: 0.461845

* Train accuracy / confusion: 79.12% / [[302, 42, 8], [46, 201, 27], [13, 31, 130]], 
* Val   accuracy / confusion: 53.85% / [[31, 13, 2], [13, 17, 5], [7, 8, 8]]

------------------------------ Epoch 146 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.394556
- Iter 024 / 025, Loss: 0.659404

* Train accuracy / confusion: 77.50% / [[300, 39, 15], [51, 192, 26], [18, 31, 128]], 
* Val   accuracy / confusion: 53.85% / [[29, 12, 5], [8, 20, 7], [6, 10, 7]]

------------------------------ Epoch 147 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.458921
- Iter 024 / 025, Loss: 0.634398

* Train accuracy / confusion: 77.75% / [[307, 39, 10], [45, 183, 40], [18, 26, 132]], 
* Val   accuracy / confusion: 57.69% / [[33, 6, 7], [6, 16, 13], [7, 5, 11]]

------------------------------ Epoch 148 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.586602
- Iter 024 / 025, Loss: 0.531084

* Train accuracy / confusion: 81.25% / [[315, 29, 12], [46, 200, 21], [14, 28, 135]], 
* Val   accuracy / confusion: 57.69% / [[30, 14, 2], [4, 22, 9], [5, 10, 8]]

------------------------------ Epoch 149 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.468155
- Iter 024 / 025, Loss: 0.530042

* Train accuracy / confusion: 79.50% / [[311, 36, 13], [55, 188, 23], [17, 20, 137]], 
* Val   accuracy / confusion: 53.85% / [[32, 11, 3], [9, 17, 9], [7, 9, 7]]

------------------------------ Epoch 150 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.614983
- Iter 024 / 025, Loss: 0.558181

* Train accuracy / confusion: 79.12% / [[311, 31, 13], [53, 194, 24], [19, 27, 128]], 
* Val   accuracy / confusion: 52.88% / [[36, 6, 4], [17, 14, 4], [9, 9, 5]]

------------------------------ Epoch 151 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.605418
- Iter 024 / 025, Loss: 0.478429

* Train accuracy / confusion: 79.75% / [[314, 42, 4], [49, 187, 27], [14, 26, 137]], 
* Val   accuracy / confusion: 55.77% / [[25, 16, 5], [6, 23, 6], [7, 6, 10]]

------------------------------ Epoch 152 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.631768
- Iter 024 / 025, Loss: 0.485120

* Train accuracy / confusion: 80.62% / [[313, 35, 9], [39, 198, 29], [19, 24, 134]], 
* Val   accuracy / confusion: 52.88% / [[28, 6, 12], [4, 18, 13], [3, 11, 9]]

------------------------------ Epoch 153 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.451216
- Iter 024 / 025, Loss: 0.513954

* Train accuracy / confusion: 81.00% / [[320, 34, 8], [43, 196, 27], [19, 21, 132]], 
* Val   accuracy / confusion: 52.88% / [[30, 8, 8], [8, 17, 10], [9, 6, 8]]

------------------------------ Epoch 154 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.395712
- Iter 024 / 025, Loss: 0.526096

* Train accuracy / confusion: 80.88% / [[309, 31, 13], [46, 202, 20], [24, 19, 136]], 
* Val   accuracy / confusion: 51.92% / [[27, 11, 8], [9, 19, 7], [5, 10, 8]]

------------------------------ Epoch 155 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.446473
- Iter 024 / 025, Loss: 0.334379

* Train accuracy / confusion: 81.88% / [[313, 28, 13], [40, 207, 23], [16, 25, 135]], 
* Val   accuracy / confusion: 56.73% / [[32, 11, 3], [7, 23, 5], [5, 14, 4]]

------------------------------ Epoch 156 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.628588
- Iter 024 / 025, Loss: 0.746246

* Train accuracy / confusion: 81.88% / [[309, 26, 17], [52, 199, 19], [9, 22, 147]], 
* Val   accuracy / confusion: 58.65% / [[36, 6, 4], [12, 15, 8], [10, 3, 10]]

------------------------------ Epoch 157 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.377932
- Iter 024 / 025, Loss: 0.560853

* Train accuracy / confusion: 81.50% / [[320, 30, 7], [44, 197, 24], [19, 24, 135]], 
* Val   accuracy / confusion: 54.81% / [[32, 7, 7], [9, 16, 10], [7, 7, 9]]

------------------------------ Epoch 158 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.557204
- Iter 024 / 025, Loss: 0.559382

* Train accuracy / confusion: 83.00% / [[321, 25, 8], [40, 205, 26], [12, 25, 138]], 
* Val   accuracy / confusion: 55.77% / [[32, 8, 6], [8, 15, 12], [7, 5, 11]]

------------------------------ Epoch 159 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.528318
- Iter 024 / 025, Loss: 0.389554

* Train accuracy / confusion: 80.88% / [[320, 29, 10], [46, 191, 31], [12, 25, 136]], 
* Val   accuracy / confusion: 52.88% / [[28, 12, 6], [8, 20, 7], [6, 10, 7]]

------------------------------ Epoch 160 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.529207
- Iter 024 / 025, Loss: 0.495277

* Train accuracy / confusion: 82.38% / [[312, 36, 10], [39, 207, 22], [11, 23, 140]], 
* Val   accuracy / confusion: 42.31% / [[27, 11, 8], [10, 10, 15], [7, 9, 7]]

------------------------------ Epoch 161 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.401975
- Iter 024 / 025, Loss: 0.594891

* Train accuracy / confusion: 83.12% / [[317, 25, 14], [44, 206, 18], [14, 20, 142]], 
* Val   accuracy / confusion: 61.54% / [[33, 11, 2], [7, 23, 5], [10, 5, 8]]

------------------------------ Epoch 162 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.200138
- Iter 024 / 025, Loss: 0.491451

* Train accuracy / confusion: 82.12% / [[317, 26, 15], [45, 208, 15], [13, 29, 132]], 
* Val   accuracy / confusion: 54.81% / [[30, 8, 8], [9, 16, 10], [5, 7, 11]]

------------------------------ Epoch 163 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.666557
- Iter 024 / 025, Loss: 0.386788

* Train accuracy / confusion: 82.38% / [[315, 31, 8], [37, 205, 26], [21, 18, 139]], 
* Val   accuracy / confusion: 54.81% / [[31, 4, 11], [7, 14, 14], [4, 7, 12]]

------------------------------ Epoch 164 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.429655
- Iter 024 / 025, Loss: 0.308347

* Train accuracy / confusion: 80.75% / [[309, 31, 14], [48, 206, 16], [19, 26, 131]], 
* Val   accuracy / confusion: 53.85% / [[29, 10, 7], [14, 17, 4], [8, 5, 10]]

------------------------------ Epoch 165 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.497100
- Iter 024 / 025, Loss: 0.395502

* Train accuracy / confusion: 78.88% / [[311, 36, 10], [53, 189, 28], [14, 28, 131]], 
* Val   accuracy / confusion: 54.81% / [[30, 13, 3], [7, 19, 9], [7, 8, 8]]

------------------------------ Epoch 166 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.453933
- Iter 024 / 025, Loss: 0.323906

* Train accuracy / confusion: 83.75% / [[320, 24, 12], [32, 212, 23], [18, 21, 138]], 
* Val   accuracy / confusion: 54.81% / [[25, 18, 3], [4, 23, 8], [5, 9, 9]]

------------------------------ Epoch 167 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.506471
- Iter 024 / 025, Loss: 0.565609

* Train accuracy / confusion: 84.00% / [[320, 23, 14], [38, 208, 21], [13, 19, 144]], 
* Val   accuracy / confusion: 50.00% / [[30, 8, 8], [13, 15, 7], [9, 7, 7]]

------------------------------ Epoch 168 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.491280
- Iter 024 / 025, Loss: 0.441340

* Train accuracy / confusion: 82.38% / [[322, 27, 7], [42, 199, 26], [16, 23, 138]], 
* Val   accuracy / confusion: 51.92% / [[30, 10, 6], [9, 15, 11], [11, 3, 9]]

------------------------------ Epoch 169 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.347183
- Iter 024 / 025, Loss: 0.411771

* Train accuracy / confusion: 83.38% / [[311, 41, 7], [35, 210, 20], [9, 21, 146]], 
* Val   accuracy / confusion: 57.69% / [[30, 15, 1], [7, 21, 7], [6, 8, 9]]

------------------------------ Epoch 170 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.553098
- Iter 024 / 025, Loss: 0.397038

* Train accuracy / confusion: 84.25% / [[320, 24, 7], [38, 210, 23], [14, 20, 144]], 
* Val   accuracy / confusion: 58.65% / [[29, 12, 5], [7, 24, 4], [4, 11, 8]]

------------------------------ Epoch 171 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.319501
- Iter 024 / 025, Loss: 0.508116

* Train accuracy / confusion: 83.25% / [[315, 28, 11], [40, 214, 17], [11, 27, 137]], 
* Val   accuracy / confusion: 53.85% / [[31, 5, 10], [13, 10, 12], [6, 2, 15]]

------------------------------ Epoch 172 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.580628
- Iter 024 / 025, Loss: 0.271479

* Train accuracy / confusion: 81.38% / [[314, 29, 11], [43, 205, 22], [21, 23, 132]], 
* Val   accuracy / confusion: 50.96% / [[32, 11, 3], [10, 16, 9], [7, 11, 5]]

------------------------------ Epoch 173 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.504495
- Iter 024 / 025, Loss: 0.449550

* Train accuracy / confusion: 84.38% / [[323, 31, 3], [35, 208, 23], [10, 23, 144]], 
* Val   accuracy / confusion: 49.04% / [[28, 13, 5], [6, 18, 11], [7, 11, 5]]

------------------------------ Epoch 174 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.437685
- Iter 024 / 025, Loss: 0.567060

* Train accuracy / confusion: 83.88% / [[325, 25, 10], [36, 208, 23], [13, 22, 138]], 
* Val   accuracy / confusion: 55.77% / [[31, 8, 7], [6, 17, 12], [6, 7, 10]]

------------------------------ Epoch 175 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.540014
- Iter 024 / 025, Loss: 0.515530

* Train accuracy / confusion: 82.38% / [[313, 28, 12], [40, 209, 20], [16, 25, 137]], 
* Val   accuracy / confusion: 55.77% / [[38, 3, 5], [13, 13, 9], [10, 6, 7]]

------------------------------ Epoch 176 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.715954
- Iter 024 / 025, Loss: 0.379334

* Train accuracy / confusion: 84.00% / [[327, 28, 6], [42, 200, 25], [12, 15, 145]], 
* Val   accuracy / confusion: 59.62% / [[33, 8, 5], [4, 23, 8], [8, 9, 6]]

------------------------------ Epoch 177 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.335396
- Iter 024 / 025, Loss: 0.373667

* Train accuracy / confusion: 83.88% / [[317, 30, 10], [42, 212, 14], [12, 21, 142]], 
* Val   accuracy / confusion: 50.00% / [[29, 10, 7], [8, 11, 16], [5, 6, 12]]

------------------------------ Epoch 178 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.671387
- Iter 024 / 025, Loss: 0.347881

* Train accuracy / confusion: 84.38% / [[321, 27, 9], [37, 206, 27], [11, 14, 148]], 
* Val   accuracy / confusion: 57.69% / [[33, 8, 5], [13, 17, 5], [7, 6, 10]]

------------------------------ Epoch 179 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.370247
- Iter 024 / 025, Loss: 0.352221

* Train accuracy / confusion: 85.38% / [[324, 24, 7], [40, 216, 16], [14, 16, 143]], 
* Val   accuracy / confusion: 50.96% / [[31, 8, 7], [11, 14, 10], [7, 8, 8]]

------------------------------ Epoch 180 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.487882
- Iter 024 / 025, Loss: 0.362547

* Train accuracy / confusion: 84.75% / [[324, 23, 9], [43, 214, 11], [15, 21, 140]], 
* Val   accuracy / confusion: 59.62% / [[33, 7, 6], [9, 19, 7], [10, 3, 10]]

------------------------------ Epoch 181 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.207010
- Iter 024 / 025, Loss: 0.544605

* Train accuracy / confusion: 87.75% / [[326, 21, 8], [33, 224, 11], [10, 15, 152]], 
* Val   accuracy / confusion: 55.77% / [[32, 9, 5], [12, 19, 4], [10, 6, 7]]

------------------------------ Epoch 182 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.273821
- Iter 024 / 025, Loss: 0.200527

* Train accuracy / confusion: 85.88% / [[325, 28, 9], [25, 217, 20], [15, 16, 145]], 
* Val   accuracy / confusion: 51.92% / [[29, 9, 8], [9, 17, 9], [6, 9, 8]]

------------------------------ Epoch 183 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.341617
- Iter 024 / 025, Loss: 0.359746

* Train accuracy / confusion: 85.00% / [[316, 26, 15], [35, 220, 13], [15, 16, 144]], 
* Val   accuracy / confusion: 48.08% / [[32, 6, 8], [12, 12, 11], [8, 9, 6]]

------------------------------ Epoch 184 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.297763
- Iter 024 / 025, Loss: 0.432673

* Train accuracy / confusion: 87.12% / [[327, 19, 6], [30, 216, 23], [9, 16, 154]], 
* Val   accuracy / confusion: 53.85% / [[28, 11, 7], [5, 23, 7], [9, 9, 5]]

------------------------------ Epoch 185 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.452935
- Iter 024 / 025, Loss: 0.415403

* Train accuracy / confusion: 84.75% / [[322, 22, 10], [37, 213, 20], [10, 23, 143]], 
* Val   accuracy / confusion: 54.81% / [[31, 12, 3], [8, 14, 13], [6, 5, 12]]

------------------------------ Epoch 186 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.437765
- Iter 024 / 025, Loss: 0.461028

* Train accuracy / confusion: 84.12% / [[329, 23, 6], [33, 210, 24], [9, 32, 134]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [9, 20, 6], [8, 9, 6]]

------------------------------ Epoch 187 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.154314
- Iter 024 / 025, Loss: 0.327644

* Train accuracy / confusion: 84.50% / [[321, 23, 12], [41, 208, 19], [10, 19, 147]], 
* Val   accuracy / confusion: 54.81% / [[29, 9, 8], [9, 17, 9], [8, 4, 11]]

------------------------------ Epoch 188 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.529080
- Iter 024 / 025, Loss: 0.496130

* Train accuracy / confusion: 85.75% / [[326, 24, 5], [38, 212, 17], [17, 13, 148]], 
* Val   accuracy / confusion: 50.96% / [[28, 13, 5], [10, 19, 6], [5, 12, 6]]

------------------------------ Epoch 189 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.492431
- Iter 024 / 025, Loss: 0.340737

* Train accuracy / confusion: 85.88% / [[325, 24, 11], [34, 220, 14], [12, 18, 142]], 
* Val   accuracy / confusion: 49.04% / [[29, 9, 8], [12, 16, 7], [13, 4, 6]]

------------------------------ Epoch 190 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.561548
- Iter 024 / 025, Loss: 0.384742

* Train accuracy / confusion: 84.75% / [[324, 26, 9], [38, 212, 17], [14, 18, 142]], 
* Val   accuracy / confusion: 50.96% / [[25, 14, 7], [5, 17, 13], [9, 3, 11]]

------------------------------ Epoch 191 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.264312
- Iter 024 / 025, Loss: 0.431335

* Train accuracy / confusion: 85.75% / [[329, 20, 7], [31, 216, 23], [10, 23, 141]], 
* Val   accuracy / confusion: 53.85% / [[32, 7, 7], [9, 17, 9], [9, 7, 7]]

------------------------------ Epoch 192 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.372741
- Iter 024 / 025, Loss: 0.399049

* Train accuracy / confusion: 83.62% / [[324, 27, 9], [35, 207, 24], [17, 19, 138]], 
* Val   accuracy / confusion: 56.73% / [[33, 9, 4], [11, 18, 6], [8, 7, 8]]

------------------------------ Epoch 193 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.363206
- Iter 024 / 025, Loss: 0.713260

* Train accuracy / confusion: 83.75% / [[315, 29, 11], [36, 215, 19], [10, 25, 140]], 
* Val   accuracy / confusion: 50.96% / [[27, 12, 7], [3, 18, 14], [5, 10, 8]]

------------------------------ Epoch 194 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.473825
- Iter 024 / 025, Loss: 0.339721

* Train accuracy / confusion: 85.50% / [[326, 24, 5], [35, 217, 19], [14, 19, 141]], 
* Val   accuracy / confusion: 51.92% / [[33, 6, 7], [10, 13, 12], [9, 6, 8]]

------------------------------ Epoch 195 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.295436
- Iter 024 / 025, Loss: 0.171286

* Train accuracy / confusion: 87.12% / [[329, 18, 9], [31, 223, 15], [16, 14, 145]], 
* Val   accuracy / confusion: 46.15% / [[27, 14, 5], [12, 13, 10], [6, 9, 8]]

------------------------------ Epoch 196 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.470868
- Iter 024 / 025, Loss: 0.260036

* Train accuracy / confusion: 86.38% / [[327, 20, 9], [30, 221, 18], [10, 22, 143]], 
* Val   accuracy / confusion: 50.00% / [[31, 10, 5], [12, 10, 13], [7, 5, 11]]

------------------------------ Epoch 197 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.261889
- Iter 024 / 025, Loss: 0.286044

* Train accuracy / confusion: 85.62% / [[320, 24, 8], [33, 218, 20], [9, 21, 147]], 
* Val   accuracy / confusion: 54.81% / [[28, 8, 10], [6, 21, 8], [5, 10, 8]]

------------------------------ Epoch 198 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.129681
- Iter 024 / 025, Loss: 0.296613

* Train accuracy / confusion: 85.12% / [[323, 26, 9], [32, 209, 26], [14, 12, 149]], 
* Val   accuracy / confusion: 54.81% / [[35, 7, 4], [10, 13, 12], [10, 4, 9]]

------------------------------ Epoch 199 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.333288
- Iter 024 / 025, Loss: 0.328183

* Train accuracy / confusion: 86.25% / [[329, 20, 8], [35, 214, 18], [14, 15, 147]], 
* Val   accuracy / confusion: 50.00% / [[28, 13, 5], [9, 17, 9], [7, 9, 7]]

------------------------------ Epoch 200 / 500, Learning rate: 4.31e-04 ------------------------------
- Iter 012 / 025, Loss: 0.384339
- Iter 024 / 025, Loss: 0.292397

* Train accuracy / confusion: 86.25% / [[325, 17, 15], [33, 219, 16], [11, 18, 146]], 
* Val   accuracy / confusion: 55.77% / [[32, 11, 3], [9, 15, 11], [8, 4, 11]]

------------------------------ Epoch 201 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.332053
- Iter 024 / 025, Loss: 0.235022

* Train accuracy / confusion: 85.62% / [[323, 24, 7], [36, 213, 21], [12, 15, 149]], 
* Val   accuracy / confusion: 65.38% / [[36, 7, 3], [10, 20, 5], [9, 2, 12]]

------------------------------ Epoch 202 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.643354
- Iter 024 / 025, Loss: 0.189164

* Train accuracy / confusion: 88.12% / [[334, 14, 5], [35, 218, 21], [8, 12, 153]], 
* Val   accuracy / confusion: 53.85% / [[33, 8, 5], [12, 15, 8], [10, 5, 8]]

------------------------------ Epoch 203 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.396244
- Iter 024 / 025, Loss: 0.211166

* Train accuracy / confusion: 88.25% / [[335, 17, 5], [33, 224, 14], [14, 11, 147]], 
* Val   accuracy / confusion: 57.69% / [[33, 7, 6], [10, 20, 5], [8, 8, 7]]

------------------------------ Epoch 204 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.382736
- Iter 024 / 025, Loss: 0.350777

* Train accuracy / confusion: 91.50% / [[340, 9, 5], [26, 231, 12], [3, 13, 161]], 
* Val   accuracy / confusion: 54.81% / [[31, 8, 7], [7, 16, 12], [8, 5, 10]]

------------------------------ Epoch 205 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.312662
- Iter 024 / 025, Loss: 0.386897

* Train accuracy / confusion: 90.75% / [[338, 16, 3], [20, 232, 14], [10, 11, 156]], 
* Val   accuracy / confusion: 57.69% / [[33, 7, 6], [4, 21, 10], [10, 7, 6]]

------------------------------ Epoch 206 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.176613
- Iter 024 / 025, Loss: 0.312661

* Train accuracy / confusion: 90.88% / [[339, 12, 4], [23, 234, 12], [6, 16, 154]], 
* Val   accuracy / confusion: 57.69% / [[31, 10, 5], [9, 20, 6], [6, 8, 9]]

------------------------------ Epoch 207 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.206107
- Iter 024 / 025, Loss: 0.201798

* Train accuracy / confusion: 90.50% / [[330, 18, 6], [20, 241, 9], [7, 16, 153]], 
* Val   accuracy / confusion: 55.77% / [[33, 6, 7], [10, 17, 8], [7, 8, 8]]

------------------------------ Epoch 208 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.146177
- Iter 024 / 025, Loss: 0.191748

* Train accuracy / confusion: 90.62% / [[341, 14, 3], [28, 229, 13], [6, 11, 155]], 
* Val   accuracy / confusion: 58.65% / [[28, 13, 5], [10, 21, 4], [5, 6, 12]]

------------------------------ Epoch 209 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.270974
- Iter 024 / 025, Loss: 0.260642

* Train accuracy / confusion: 90.38% / [[335, 11, 6], [19, 236, 18], [6, 17, 152]], 
* Val   accuracy / confusion: 62.50% / [[32, 8, 6], [6, 25, 4], [11, 4, 8]]

------------------------------ Epoch 210 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.273629
- Iter 024 / 025, Loss: 0.312531

* Train accuracy / confusion: 87.75% / [[334, 14, 6], [26, 225, 18], [14, 20, 143]], 
* Val   accuracy / confusion: 53.85% / [[31, 10, 5], [8, 17, 10], [9, 6, 8]]

------------------------------ Epoch 211 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.335792
- Iter 024 / 025, Loss: 0.335360

* Train accuracy / confusion: 90.75% / [[334, 17, 4], [27, 238, 6], [6, 14, 154]], 
* Val   accuracy / confusion: 50.00% / [[24, 12, 10], [5, 19, 11], [6, 8, 9]]

------------------------------ Epoch 212 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.264797
- Iter 024 / 025, Loss: 0.216431

* Train accuracy / confusion: 91.75% / [[333, 13, 7], [22, 241, 7], [8, 9, 160]], 
* Val   accuracy / confusion: 55.77% / [[31, 10, 5], [10, 20, 5], [6, 10, 7]]

------------------------------ Epoch 213 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.226490
- Iter 024 / 025, Loss: 0.236492

* Train accuracy / confusion: 90.25% / [[335, 15, 5], [23, 232, 15], [7, 13, 155]], 
* Val   accuracy / confusion: 56.73% / [[31, 7, 8], [9, 18, 8], [8, 5, 10]]

------------------------------ Epoch 214 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.165873
- Iter 024 / 025, Loss: 0.274816

* Train accuracy / confusion: 91.00% / [[338, 13, 6], [30, 231, 6], [8, 9, 159]], 
* Val   accuracy / confusion: 57.69% / [[32, 9, 5], [8, 20, 7], [4, 11, 8]]

------------------------------ Epoch 215 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.517417
- Iter 024 / 025, Loss: 0.257123

* Train accuracy / confusion: 92.00% / [[346, 9, 3], [18, 237, 13], [10, 11, 153]], 
* Val   accuracy / confusion: 52.88% / [[29, 9, 8], [7, 14, 14], [4, 7, 12]]

------------------------------ Epoch 216 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.278361
- Iter 024 / 025, Loss: 0.145732

* Train accuracy / confusion: 92.25% / [[337, 17, 3], [16, 242, 8], [7, 11, 159]], 
* Val   accuracy / confusion: 54.81% / [[32, 9, 5], [11, 15, 9], [7, 6, 10]]

------------------------------ Epoch 217 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.250759
- Iter 024 / 025, Loss: 0.223152

* Train accuracy / confusion: 90.50% / [[337, 14, 6], [23, 230, 15], [6, 12, 157]], 
* Val   accuracy / confusion: 50.96% / [[29, 13, 4], [9, 17, 9], [9, 7, 7]]

------------------------------ Epoch 218 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.216346
- Iter 024 / 025, Loss: 0.184037

* Train accuracy / confusion: 92.25% / [[346, 10, 3], [16, 234, 16], [7, 10, 158]], 
* Val   accuracy / confusion: 51.92% / [[29, 11, 6], [8, 16, 11], [8, 6, 9]]

------------------------------ Epoch 219 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.165895
- Iter 024 / 025, Loss: 0.259208

* Train accuracy / confusion: 91.88% / [[331, 15, 3], [27, 243, 4], [8, 8, 161]], 
* Val   accuracy / confusion: 52.88% / [[31, 10, 5], [7, 15, 13], [10, 4, 9]]

------------------------------ Epoch 220 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.168059
- Iter 024 / 025, Loss: 0.112269

* Train accuracy / confusion: 90.50% / [[333, 18, 5], [25, 230, 10], [8, 10, 161]], 
* Val   accuracy / confusion: 53.85% / [[30, 9, 7], [5, 18, 12], [6, 9, 8]]

------------------------------ Epoch 221 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.337777
- Iter 024 / 025, Loss: 0.144238

* Train accuracy / confusion: 90.88% / [[334, 13, 8], [23, 233, 11], [6, 12, 160]], 
* Val   accuracy / confusion: 56.73% / [[29, 11, 6], [5, 22, 8], [8, 7, 8]]

------------------------------ Epoch 222 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.185399
- Iter 024 / 025, Loss: 0.213856

* Train accuracy / confusion: 90.12% / [[340, 11, 6], [23, 228, 18], [3, 18, 153]], 
* Val   accuracy / confusion: 55.77% / [[33, 8, 5], [9, 16, 10], [4, 10, 9]]

------------------------------ Epoch 223 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.207874
- Iter 024 / 025, Loss: 0.236908

* Train accuracy / confusion: 91.62% / [[332, 18, 5], [23, 235, 8], [6, 7, 166]], 
* Val   accuracy / confusion: 58.65% / [[33, 6, 7], [6, 20, 9], [11, 4, 8]]

------------------------------ Epoch 224 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.105393
- Iter 024 / 025, Loss: 0.099549

* Train accuracy / confusion: 91.88% / [[342, 9, 3], [23, 234, 10], [4, 16, 159]], 
* Val   accuracy / confusion: 57.69% / [[31, 9, 6], [6, 24, 5], [8, 10, 5]]

------------------------------ Epoch 225 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.113747
- Iter 024 / 025, Loss: 0.241698

* Train accuracy / confusion: 90.25% / [[331, 20, 6], [26, 231, 10], [5, 11, 160]], 
* Val   accuracy / confusion: 48.08% / [[30, 9, 7], [14, 12, 9], [8, 7, 8]]

------------------------------ Epoch 226 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.114620
- Iter 024 / 025, Loss: 0.157365

* Train accuracy / confusion: 93.50% / [[346, 5, 1], [22, 241, 9], [7, 8, 161]], 
* Val   accuracy / confusion: 53.85% / [[33, 7, 6], [8, 17, 10], [7, 10, 6]]

------------------------------ Epoch 227 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.149773
- Iter 024 / 025, Loss: 0.173532

* Train accuracy / confusion: 92.88% / [[339, 8, 6], [19, 242, 10], [5, 9, 162]], 
* Val   accuracy / confusion: 52.88% / [[28, 15, 3], [8, 21, 6], [12, 5, 6]]

------------------------------ Epoch 228 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.180852
- Iter 024 / 025, Loss: 0.455283

* Train accuracy / confusion: 90.75% / [[337, 15, 2], [22, 232, 18], [10, 7, 157]], 
* Val   accuracy / confusion: 60.58% / [[32, 7, 7], [4, 23, 8], [8, 7, 8]]

------------------------------ Epoch 229 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.146494
- Iter 024 / 025, Loss: 0.272966

* Train accuracy / confusion: 93.25% / [[339, 10, 7], [15, 250, 3], [11, 8, 157]], 
* Val   accuracy / confusion: 55.77% / [[33, 6, 7], [5, 19, 11], [7, 10, 6]]

------------------------------ Epoch 230 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.216055
- Iter 024 / 025, Loss: 0.244695

* Train accuracy / confusion: 89.50% / [[337, 18, 3], [28, 228, 11], [8, 16, 151]], 
* Val   accuracy / confusion: 46.15% / [[27, 8, 11], [9, 15, 11], [10, 7, 6]]

------------------------------ Epoch 231 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.288351
- Iter 024 / 025, Loss: 0.208135

* Train accuracy / confusion: 92.25% / [[347, 8, 2], [21, 234, 9], [9, 13, 157]], 
* Val   accuracy / confusion: 55.77% / [[31, 11, 4], [5, 18, 12], [5, 9, 9]]

------------------------------ Epoch 232 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.183412
- Iter 024 / 025, Loss: 0.171689

* Train accuracy / confusion: 93.00% / [[339, 14, 2], [19, 243, 6], [8, 7, 162]], 
* Val   accuracy / confusion: 52.88% / [[26, 10, 10], [5, 19, 11], [8, 5, 10]]

------------------------------ Epoch 233 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.110824
- Iter 024 / 025, Loss: 0.363339

* Train accuracy / confusion: 93.00% / [[343, 10, 1], [21, 240, 7], [6, 11, 161]], 
* Val   accuracy / confusion: 58.65% / [[33, 10, 3], [9, 19, 7], [6, 8, 9]]

------------------------------ Epoch 234 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.169424
- Iter 024 / 025, Loss: 0.200980

* Train accuracy / confusion: 91.75% / [[333, 16, 3], [20, 242, 8], [8, 11, 159]], 
* Val   accuracy / confusion: 57.69% / [[32, 6, 8], [8, 18, 9], [3, 10, 10]]

------------------------------ Epoch 235 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.244287
- Iter 024 / 025, Loss: 0.203458

* Train accuracy / confusion: 92.50% / [[341, 12, 5], [17, 242, 9], [9, 8, 157]], 
* Val   accuracy / confusion: 53.85% / [[31, 9, 6], [10, 17, 8], [12, 3, 8]]

------------------------------ Epoch 236 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.238133
- Iter 024 / 025, Loss: 0.184420

* Train accuracy / confusion: 92.62% / [[343, 13, 3], [16, 238, 12], [4, 11, 160]], 
* Val   accuracy / confusion: 62.50% / [[36, 4, 6], [6, 21, 8], [8, 7, 8]]

------------------------------ Epoch 237 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.256160
- Iter 024 / 025, Loss: 0.180827

* Train accuracy / confusion: 91.25% / [[342, 11, 8], [23, 230, 13], [2, 13, 158]], 
* Val   accuracy / confusion: 47.12% / [[28, 14, 4], [7, 12, 16], [5, 9, 9]]

------------------------------ Epoch 238 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.285321
- Iter 024 / 025, Loss: 0.293234

* Train accuracy / confusion: 92.75% / [[344, 13, 1], [18, 239, 9], [6, 11, 159]], 
* Val   accuracy / confusion: 53.85% / [[32, 6, 8], [9, 17, 9], [11, 5, 7]]

------------------------------ Epoch 239 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.086342
- Iter 024 / 025, Loss: 0.215297

* Train accuracy / confusion: 92.38% / [[337, 12, 7], [13, 241, 12], [6, 11, 161]], 
* Val   accuracy / confusion: 51.92% / [[30, 11, 5], [13, 14, 8], [7, 6, 10]]

------------------------------ Epoch 240 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.222406
- Iter 024 / 025, Loss: 0.323228

* Train accuracy / confusion: 93.25% / [[337, 14, 2], [18, 241, 10], [4, 6, 168]], 
* Val   accuracy / confusion: 56.73% / [[32, 10, 4], [6, 20, 9], [6, 10, 7]]

------------------------------ Epoch 241 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.239512
- Iter 024 / 025, Loss: 0.202045

* Train accuracy / confusion: 93.38% / [[344, 6, 6], [16, 243, 10], [6, 9, 160]], 
* Val   accuracy / confusion: 55.77% / [[31, 8, 7], [8, 19, 8], [8, 7, 8]]

------------------------------ Epoch 242 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.253574
- Iter 024 / 025, Loss: 0.167748

* Train accuracy / confusion: 92.38% / [[341, 13, 2], [23, 236, 10], [4, 9, 162]], 
* Val   accuracy / confusion: 53.85% / [[34, 9, 3], [11, 15, 9], [8, 8, 7]]

------------------------------ Epoch 243 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.192399
- Iter 024 / 025, Loss: 0.202231

* Train accuracy / confusion: 93.25% / [[337, 16, 2], [15, 250, 8], [6, 7, 159]], 
* Val   accuracy / confusion: 54.81% / [[29, 14, 3], [6, 17, 12], [7, 5, 11]]

------------------------------ Epoch 244 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.233790
- Iter 024 / 025, Loss: 0.110893

* Train accuracy / confusion: 92.25% / [[344, 11, 2], [23, 234, 11], [3, 12, 160]], 
* Val   accuracy / confusion: 56.73% / [[28, 8, 10], [6, 19, 10], [5, 6, 12]]

------------------------------ Epoch 245 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.215096
- Iter 024 / 025, Loss: 0.196891

* Train accuracy / confusion: 93.50% / [[343, 14, 3], [12, 243, 12], [5, 6, 162]], 
* Val   accuracy / confusion: 56.73% / [[32, 6, 8], [8, 16, 11], [6, 6, 11]]

------------------------------ Epoch 246 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.234050
- Iter 024 / 025, Loss: 0.294276

* Train accuracy / confusion: 91.50% / [[337, 17, 4], [17, 238, 10], [7, 13, 157]], 
* Val   accuracy / confusion: 52.88% / [[26, 12, 8], [6, 17, 12], [5, 6, 12]]

------------------------------ Epoch 247 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.228788
- Iter 024 / 025, Loss: 0.285270

* Train accuracy / confusion: 92.75% / [[339, 11, 5], [17, 243, 7], [9, 9, 160]], 
* Val   accuracy / confusion: 54.81% / [[33, 5, 8], [5, 19, 11], [8, 10, 5]]

------------------------------ Epoch 248 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.236927
- Iter 024 / 025, Loss: 0.194348

* Train accuracy / confusion: 93.62% / [[343, 17, 1], [11, 245, 10], [6, 6, 161]], 
* Val   accuracy / confusion: 53.85% / [[28, 10, 8], [6, 21, 8], [8, 8, 7]]

------------------------------ Epoch 249 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.181258
- Iter 024 / 025, Loss: 0.259377

* Train accuracy / confusion: 92.62% / [[335, 13, 4], [17, 247, 10], [4, 11, 159]], 
* Val   accuracy / confusion: 50.96% / [[28, 13, 5], [9, 17, 9], [9, 6, 8]]

------------------------------ Epoch 250 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.139363
- Iter 024 / 025, Loss: 0.135876

* Train accuracy / confusion: 92.38% / [[347, 7, 3], [21, 232, 16], [7, 7, 160]], 
* Val   accuracy / confusion: 49.04% / [[27, 10, 9], [8, 18, 9], [10, 7, 6]]

------------------------------ Epoch 251 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.096167
- Iter 024 / 025, Loss: 0.417475

* Train accuracy / confusion: 93.25% / [[338, 14, 1], [19, 245, 7], [7, 6, 163]], 
* Val   accuracy / confusion: 52.88% / [[25, 13, 8], [6, 19, 10], [6, 6, 11]]

------------------------------ Epoch 252 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.421779
- Iter 024 / 025, Loss: 0.196150

* Train accuracy / confusion: 93.88% / [[343, 9, 5], [16, 250, 2], [8, 9, 158]], 
* Val   accuracy / confusion: 57.69% / [[29, 10, 7], [7, 19, 9], [6, 5, 12]]

------------------------------ Epoch 253 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.112143
- Iter 024 / 025, Loss: 0.267457

* Train accuracy / confusion: 93.50% / [[350, 8, 1], [12, 239, 17], [8, 6, 159]], 
* Val   accuracy / confusion: 56.73% / [[31, 8, 7], [5, 21, 9], [9, 7, 7]]

------------------------------ Epoch 254 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.188120
- Iter 024 / 025, Loss: 0.280496

* Train accuracy / confusion: 93.75% / [[342, 13, 4], [13, 245, 7], [2, 11, 163]], 
* Val   accuracy / confusion: 55.77% / [[29, 10, 7], [5, 21, 9], [8, 7, 8]]

------------------------------ Epoch 255 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.143422
- Iter 024 / 025, Loss: 0.095345

* Train accuracy / confusion: 92.50% / [[345, 10, 2], [19, 243, 9], [8, 12, 152]], 
* Val   accuracy / confusion: 56.73% / [[33, 8, 5], [9, 17, 9], [8, 6, 9]]

------------------------------ Epoch 256 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.117716
- Iter 024 / 025, Loss: 0.192805

* Train accuracy / confusion: 95.38% / [[350, 5, 1], [11, 250, 8], [2, 10, 163]], 
* Val   accuracy / confusion: 52.88% / [[26, 13, 7], [8, 22, 5], [7, 9, 7]]

------------------------------ Epoch 257 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.107953
- Iter 024 / 025, Loss: 0.378242

* Train accuracy / confusion: 91.50% / [[340, 13, 3], [22, 238, 9], [9, 12, 154]], 
* Val   accuracy / confusion: 61.54% / [[33, 5, 8], [8, 20, 7], [7, 5, 11]]

------------------------------ Epoch 258 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.317979
- Iter 024 / 025, Loss: 0.212161

* Train accuracy / confusion: 93.00% / [[339, 9, 8], [10, 247, 10], [5, 14, 158]], 
* Val   accuracy / confusion: 54.81% / [[29, 14, 3], [7, 18, 10], [5, 8, 10]]

------------------------------ Epoch 259 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.152149
- Iter 024 / 025, Loss: 0.276231

* Train accuracy / confusion: 93.25% / [[341, 13, 3], [15, 244, 6], [8, 9, 161]], 
* Val   accuracy / confusion: 55.77% / [[34, 6, 6], [10, 16, 9], [12, 3, 8]]

------------------------------ Epoch 260 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.157265
- Iter 024 / 025, Loss: 0.147600

* Train accuracy / confusion: 92.88% / [[339, 11, 4], [23, 240, 9], [4, 6, 164]], 
* Val   accuracy / confusion: 57.69% / [[29, 11, 6], [7, 20, 8], [6, 6, 11]]

------------------------------ Epoch 261 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.077080
- Iter 024 / 025, Loss: 0.224121

* Train accuracy / confusion: 93.38% / [[344, 7, 5], [16, 240, 9], [7, 9, 163]], 
* Val   accuracy / confusion: 56.73% / [[31, 11, 4], [7, 18, 10], [6, 7, 10]]

------------------------------ Epoch 262 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.070018
- Iter 024 / 025, Loss: 0.266640

* Train accuracy / confusion: 93.38% / [[342, 5, 10], [13, 244, 9], [5, 11, 161]], 
* Val   accuracy / confusion: 53.85% / [[26, 11, 9], [6, 22, 7], [7, 8, 8]]

------------------------------ Epoch 263 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.221576
- Iter 024 / 025, Loss: 0.071263

* Train accuracy / confusion: 93.75% / [[343, 12, 1], [18, 245, 6], [6, 7, 162]], 
* Val   accuracy / confusion: 64.42% / [[35, 5, 6], [4, 24, 7], [6, 9, 8]]

------------------------------ Epoch 264 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.297352
- Iter 024 / 025, Loss: 0.206362

* Train accuracy / confusion: 93.50% / [[339, 16, 2], [16, 248, 5], [5, 8, 161]], 
* Val   accuracy / confusion: 50.00% / [[30, 10, 6], [11, 13, 11], [10, 4, 9]]

------------------------------ Epoch 265 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.119499
- Iter 024 / 025, Loss: 0.099981

* Train accuracy / confusion: 92.75% / [[338, 14, 2], [17, 244, 9], [9, 7, 160]], 
* Val   accuracy / confusion: 47.12% / [[29, 11, 6], [11, 14, 10], [11, 6, 6]]

------------------------------ Epoch 266 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.228124
- Iter 024 / 025, Loss: 0.083309

* Train accuracy / confusion: 93.12% / [[335, 13, 6], [20, 244, 6], [1, 9, 166]], 
* Val   accuracy / confusion: 53.85% / [[31, 9, 6], [11, 18, 6], [8, 8, 7]]

------------------------------ Epoch 267 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.237148
- Iter 024 / 025, Loss: 0.348732

* Train accuracy / confusion: 93.50% / [[349, 5, 3], [13, 241, 16], [4, 11, 158]], 
* Val   accuracy / confusion: 51.92% / [[32, 7, 7], [13, 16, 6], [10, 7, 6]]

------------------------------ Epoch 268 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.316360
- Iter 024 / 025, Loss: 0.133791

* Train accuracy / confusion: 93.38% / [[343, 10, 2], [17, 247, 6], [9, 9, 157]], 
* Val   accuracy / confusion: 56.73% / [[28, 11, 7], [6, 22, 7], [6, 8, 9]]

------------------------------ Epoch 269 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.199266
- Iter 024 / 025, Loss: 0.104278

* Train accuracy / confusion: 93.50% / [[339, 14, 4], [16, 242, 6], [6, 6, 167]], 
* Val   accuracy / confusion: 58.65% / [[33, 8, 5], [8, 20, 7], [5, 10, 8]]

------------------------------ Epoch 270 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.176051
- Iter 024 / 025, Loss: 0.284246

* Train accuracy / confusion: 94.00% / [[350, 8, 4], [13, 248, 8], [5, 10, 154]], 
* Val   accuracy / confusion: 50.00% / [[28, 11, 7], [9, 14, 12], [6, 7, 10]]

------------------------------ Epoch 271 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.079955
- Iter 024 / 025, Loss: 0.222130

* Train accuracy / confusion: 94.75% / [[347, 8, 1], [13, 248, 8], [4, 8, 163]], 
* Val   accuracy / confusion: 54.81% / [[30, 9, 7], [10, 17, 8], [7, 6, 10]]

------------------------------ Epoch 272 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.249060
- Iter 024 / 025, Loss: 0.168707

* Train accuracy / confusion: 93.62% / [[338, 13, 3], [18, 243, 7], [5, 5, 168]], 
* Val   accuracy / confusion: 56.73% / [[31, 7, 8], [7, 20, 8], [11, 4, 8]]

------------------------------ Epoch 273 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.259684
- Iter 024 / 025, Loss: 0.207196

* Train accuracy / confusion: 92.50% / [[344, 11, 1], [17, 236, 13], [7, 11, 160]], 
* Val   accuracy / confusion: 51.92% / [[31, 9, 6], [8, 16, 11], [8, 8, 7]]

------------------------------ Epoch 274 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.127161
- Iter 024 / 025, Loss: 0.173116

* Train accuracy / confusion: 93.88% / [[341, 11, 5], [16, 247, 5], [5, 7, 163]], 
* Val   accuracy / confusion: 61.54% / [[34, 9, 3], [9, 23, 3], [9, 7, 7]]

------------------------------ Epoch 275 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.124455
- Iter 024 / 025, Loss: 0.190540

* Train accuracy / confusion: 94.00% / [[341, 14, 2], [14, 249, 6], [3, 9, 162]], 
* Val   accuracy / confusion: 46.15% / [[28, 14, 4], [9, 14, 12], [5, 12, 6]]

------------------------------ Epoch 276 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.111439
- Iter 024 / 025, Loss: 0.269052

* Train accuracy / confusion: 93.00% / [[337, 14, 4], [15, 241, 13], [2, 8, 166]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [8, 17, 10], [9, 6, 8]]

------------------------------ Epoch 277 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.120613
- Iter 024 / 025, Loss: 0.338537

* Train accuracy / confusion: 94.50% / [[347, 9, 3], [14, 244, 7], [4, 7, 165]], 
* Val   accuracy / confusion: 52.88% / [[28, 12, 6], [7, 18, 10], [6, 8, 9]]

------------------------------ Epoch 278 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.194567
- Iter 024 / 025, Loss: 0.242922

* Train accuracy / confusion: 93.75% / [[345, 12, 0], [16, 241, 11], [5, 6, 164]], 
* Val   accuracy / confusion: 51.92% / [[28, 11, 7], [11, 17, 7], [8, 6, 9]]

------------------------------ Epoch 279 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.121854
- Iter 024 / 025, Loss: 0.266573

* Train accuracy / confusion: 94.25% / [[344, 10, 5], [15, 243, 8], [2, 6, 167]], 
* Val   accuracy / confusion: 54.81% / [[30, 10, 6], [6, 19, 10], [9, 6, 8]]

------------------------------ Epoch 280 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.257200
- Iter 024 / 025, Loss: 0.517809

* Train accuracy / confusion: 91.75% / [[338, 10, 6], [27, 232, 11], [6, 6, 164]], 
* Val   accuracy / confusion: 54.81% / [[31, 7, 8], [9, 17, 9], [6, 8, 9]]

------------------------------ Epoch 281 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.135647
- Iter 024 / 025, Loss: 0.299728

* Train accuracy / confusion: 94.25% / [[340, 10, 6], [13, 247, 8], [4, 5, 167]], 
* Val   accuracy / confusion: 51.92% / [[27, 11, 8], [7, 16, 12], [5, 7, 11]]

------------------------------ Epoch 282 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.191543
- Iter 024 / 025, Loss: 0.100070

* Train accuracy / confusion: 93.50% / [[347, 9, 2], [17, 243, 4], [7, 13, 158]], 
* Val   accuracy / confusion: 58.65% / [[29, 9, 8], [4, 21, 10], [4, 8, 11]]

------------------------------ Epoch 283 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.132695
- Iter 024 / 025, Loss: 0.181023

* Train accuracy / confusion: 94.50% / [[337, 15, 3], [12, 248, 8], [2, 4, 171]], 
* Val   accuracy / confusion: 50.96% / [[29, 11, 6], [10, 20, 5], [8, 11, 4]]

------------------------------ Epoch 284 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.270355
- Iter 024 / 025, Loss: 0.077932

* Train accuracy / confusion: 94.12% / [[338, 16, 1], [18, 248, 3], [2, 7, 167]], 
* Val   accuracy / confusion: 52.88% / [[27, 13, 6], [9, 17, 9], [8, 4, 11]]

------------------------------ Epoch 285 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.222773
- Iter 024 / 025, Loss: 0.281830

* Train accuracy / confusion: 94.75% / [[345, 9, 6], [13, 246, 8], [0, 6, 167]], 
* Val   accuracy / confusion: 57.69% / [[32, 9, 5], [9, 20, 6], [5, 10, 8]]

------------------------------ Epoch 286 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.109784
- Iter 024 / 025, Loss: 0.178376

* Train accuracy / confusion: 94.75% / [[350, 4, 3], [12, 246, 7], [4, 12, 162]], 
* Val   accuracy / confusion: 55.77% / [[29, 13, 4], [5, 20, 10], [6, 8, 9]]

------------------------------ Epoch 287 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.183820
- Iter 024 / 025, Loss: 0.267102

* Train accuracy / confusion: 93.50% / [[344, 7, 2], [17, 250, 7], [8, 11, 154]], 
* Val   accuracy / confusion: 46.15% / [[23, 16, 7], [7, 17, 11], [11, 4, 8]]

------------------------------ Epoch 288 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.063390
- Iter 024 / 025, Loss: 0.128530

* Train accuracy / confusion: 94.62% / [[344, 11, 8], [9, 249, 8], [2, 5, 164]], 
* Val   accuracy / confusion: 50.00% / [[32, 6, 8], [8, 15, 12], [8, 10, 5]]

------------------------------ Epoch 289 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.340220
- Iter 024 / 025, Loss: 0.066165

* Train accuracy / confusion: 95.12% / [[347, 7, 2], [16, 243, 6], [3, 5, 171]], 
* Val   accuracy / confusion: 42.31% / [[25, 14, 7], [12, 11, 12], [7, 8, 8]]

------------------------------ Epoch 290 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.181547
- Iter 024 / 025, Loss: 0.223511

* Train accuracy / confusion: 93.50% / [[345, 7, 3], [23, 237, 7], [3, 9, 166]], 
* Val   accuracy / confusion: 49.04% / [[28, 11, 7], [10, 16, 9], [7, 9, 7]]

------------------------------ Epoch 291 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.268413
- Iter 024 / 025, Loss: 0.132595

* Train accuracy / confusion: 95.25% / [[342, 11, 0], [11, 255, 8], [4, 4, 165]], 
* Val   accuracy / confusion: 54.81% / [[31, 11, 4], [7, 18, 10], [7, 8, 8]]

------------------------------ Epoch 292 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.093425
- Iter 024 / 025, Loss: 0.151983

* Train accuracy / confusion: 93.75% / [[348, 5, 3], [17, 239, 9], [6, 10, 163]], 
* Val   accuracy / confusion: 59.62% / [[32, 9, 5], [7, 20, 8], [7, 6, 10]]

------------------------------ Epoch 293 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.159585
- Iter 024 / 025, Loss: 0.078683

* Train accuracy / confusion: 93.62% / [[342, 9, 4], [17, 244, 8], [2, 11, 163]], 
* Val   accuracy / confusion: 60.58% / [[33, 9, 4], [10, 20, 5], [6, 7, 10]]

------------------------------ Epoch 294 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.211762
- Iter 024 / 025, Loss: 0.138298

* Train accuracy / confusion: 93.38% / [[342, 8, 7], [13, 249, 10], [8, 7, 156]], 
* Val   accuracy / confusion: 54.81% / [[27, 10, 9], [6, 19, 10], [7, 5, 11]]

------------------------------ Epoch 295 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.299497
- Iter 024 / 025, Loss: 0.180267

* Train accuracy / confusion: 93.38% / [[345, 11, 2], [23, 237, 7], [4, 6, 165]], 
* Val   accuracy / confusion: 50.96% / [[29, 10, 7], [7, 17, 11], [9, 7, 7]]

------------------------------ Epoch 296 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.172891
- Iter 024 / 025, Loss: 0.167298

* Train accuracy / confusion: 94.62% / [[342, 10, 4], [13, 250, 9], [2, 5, 165]], 
* Val   accuracy / confusion: 55.77% / [[32, 7, 7], [10, 15, 10], [9, 3, 11]]

------------------------------ Epoch 297 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.131589
- Iter 024 / 025, Loss: 0.081312

* Train accuracy / confusion: 93.38% / [[343, 10, 4], [17, 241, 8], [7, 7, 163]], 
* Val   accuracy / confusion: 50.00% / [[31, 9, 6], [12, 13, 10], [8, 7, 8]]

------------------------------ Epoch 298 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.148117
- Iter 024 / 025, Loss: 0.166932

* Train accuracy / confusion: 95.62% / [[349, 6, 1], [15, 252, 4], [5, 4, 164]], 
* Val   accuracy / confusion: 54.81% / [[29, 12, 5], [9, 19, 7], [7, 7, 9]]

------------------------------ Epoch 299 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.151283
- Iter 024 / 025, Loss: 0.236551

* Train accuracy / confusion: 95.50% / [[344, 8, 4], [5, 251, 9], [1, 9, 169]], 
* Val   accuracy / confusion: 56.73% / [[32, 8, 6], [10, 17, 8], [8, 5, 10]]

------------------------------ Epoch 300 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.074321
- Iter 024 / 025, Loss: 0.107504

* Train accuracy / confusion: 94.38% / [[344, 7, 4], [11, 251, 8], [4, 11, 160]], 
* Val   accuracy / confusion: 54.81% / [[30, 9, 7], [6, 20, 9], [8, 8, 7]]

------------------------------ Epoch 301 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.159379
- Iter 024 / 025, Loss: 0.174928

* Train accuracy / confusion: 94.00% / [[347, 8, 0], [18, 241, 7], [5, 10, 164]], 
* Val   accuracy / confusion: 56.73% / [[32, 6, 8], [9, 17, 9], [7, 6, 10]]

------------------------------ Epoch 302 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.280453
- Iter 024 / 025, Loss: 0.155018

* Train accuracy / confusion: 95.88% / [[347, 4, 4], [12, 256, 2], [3, 8, 164]], 
* Val   accuracy / confusion: 50.00% / [[29, 10, 7], [8, 19, 8], [10, 9, 4]]

------------------------------ Epoch 303 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.106198
- Iter 024 / 025, Loss: 0.150307

* Train accuracy / confusion: 94.25% / [[344, 9, 3], [16, 243, 9], [1, 8, 167]], 
* Val   accuracy / confusion: 56.73% / [[32, 5, 9], [9, 18, 8], [6, 8, 9]]

------------------------------ Epoch 304 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.345803
- Iter 024 / 025, Loss: 0.086391

* Train accuracy / confusion: 94.50% / [[343, 10, 1], [8, 251, 10], [5, 10, 162]], 
* Val   accuracy / confusion: 54.81% / [[30, 10, 6], [7, 20, 8], [7, 9, 7]]

------------------------------ Epoch 305 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.210816
- Iter 024 / 025, Loss: 0.365853

* Train accuracy / confusion: 93.25% / [[338, 20, 4], [14, 252, 3], [5, 8, 156]], 
* Val   accuracy / confusion: 50.96% / [[29, 10, 7], [8, 17, 10], [4, 12, 7]]

------------------------------ Epoch 306 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.122314
- Iter 024 / 025, Loss: 0.120359

* Train accuracy / confusion: 95.88% / [[350, 9, 2], [4, 253, 6], [3, 9, 164]], 
* Val   accuracy / confusion: 54.81% / [[32, 6, 8], [5, 19, 11], [11, 6, 6]]

------------------------------ Epoch 307 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.234268
- Iter 024 / 025, Loss: 0.176610

* Train accuracy / confusion: 95.12% / [[347, 11, 3], [7, 252, 8], [2, 8, 162]], 
* Val   accuracy / confusion: 51.92% / [[32, 9, 5], [9, 16, 10], [7, 10, 6]]

------------------------------ Epoch 308 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.102975
- Iter 024 / 025, Loss: 0.116033

* Train accuracy / confusion: 95.12% / [[349, 7, 2], [14, 248, 4], [5, 7, 164]], 
* Val   accuracy / confusion: 51.92% / [[31, 8, 7], [12, 12, 11], [7, 5, 11]]

------------------------------ Epoch 309 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.122862
- Iter 024 / 025, Loss: 0.300076

* Train accuracy / confusion: 94.25% / [[344, 9, 3], [14, 246, 8], [4, 8, 164]], 
* Val   accuracy / confusion: 50.00% / [[31, 8, 7], [9, 16, 10], [8, 10, 5]]

------------------------------ Epoch 310 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.122977
- Iter 024 / 025, Loss: 0.461867

* Train accuracy / confusion: 94.00% / [[342, 12, 3], [7, 245, 13], [5, 8, 165]], 
* Val   accuracy / confusion: 53.85% / [[27, 12, 7], [7, 19, 9], [6, 7, 10]]

------------------------------ Epoch 311 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.227452
- Iter 024 / 025, Loss: 0.176377

* Train accuracy / confusion: 93.88% / [[352, 6, 0], [18, 238, 12], [4, 9, 161]], 
* Val   accuracy / confusion: 58.65% / [[31, 9, 6], [6, 21, 8], [6, 8, 9]]

------------------------------ Epoch 312 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.140554
- Iter 024 / 025, Loss: 0.115385

* Train accuracy / confusion: 94.38% / [[340, 8, 6], [17, 253, 5], [2, 7, 162]], 
* Val   accuracy / confusion: 52.88% / [[29, 11, 6], [7, 18, 10], [8, 7, 8]]

------------------------------ Epoch 313 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.295228
- Iter 024 / 025, Loss: 0.451819

* Train accuracy / confusion: 92.88% / [[333, 15, 4], [16, 245, 10], [7, 5, 165]], 
* Val   accuracy / confusion: 50.96% / [[29, 11, 6], [9, 17, 9], [7, 9, 7]]

------------------------------ Epoch 314 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.103954
- Iter 024 / 025, Loss: 0.069360

* Train accuracy / confusion: 95.25% / [[340, 7, 1], [10, 256, 8], [4, 8, 166]], 
* Val   accuracy / confusion: 55.77% / [[33, 6, 7], [9, 17, 9], [6, 9, 8]]

------------------------------ Epoch 315 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.146370
- Iter 024 / 025, Loss: 0.136412

* Train accuracy / confusion: 93.88% / [[342, 9, 4], [16, 247, 7], [5, 8, 162]], 
* Val   accuracy / confusion: 53.85% / [[31, 10, 5], [6, 19, 10], [11, 6, 6]]

------------------------------ Epoch 316 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.109177
- Iter 024 / 025, Loss: 0.108917

* Train accuracy / confusion: 95.00% / [[347, 9, 1], [13, 248, 7], [3, 7, 165]], 
* Val   accuracy / confusion: 51.92% / [[29, 9, 8], [8, 17, 10], [9, 6, 8]]

------------------------------ Epoch 317 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.144865
- Iter 024 / 025, Loss: 0.135217

* Train accuracy / confusion: 95.00% / [[340, 10, 5], [11, 251, 4], [3, 7, 169]], 
* Val   accuracy / confusion: 50.96% / [[24, 14, 8], [5, 19, 11], [7, 6, 10]]

------------------------------ Epoch 318 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.267748
- Iter 024 / 025, Loss: 0.219531

* Train accuracy / confusion: 94.38% / [[352, 10, 1], [12, 243, 7], [5, 10, 160]], 
* Val   accuracy / confusion: 49.04% / [[29, 10, 7], [7, 15, 13], [9, 7, 7]]

------------------------------ Epoch 319 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.157588
- Iter 024 / 025, Loss: 0.097177

* Train accuracy / confusion: 95.38% / [[345, 8, 0], [14, 252, 6], [4, 5, 166]], 
* Val   accuracy / confusion: 54.81% / [[30, 10, 6], [8, 17, 10], [7, 6, 10]]

------------------------------ Epoch 320 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.264444
- Iter 024 / 025, Loss: 0.130986

* Train accuracy / confusion: 94.75% / [[344, 11, 1], [13, 249, 5], [6, 6, 165]], 
* Val   accuracy / confusion: 62.50% / [[34, 5, 7], [7, 22, 6], [7, 7, 9]]

------------------------------ Epoch 321 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.196131
- Iter 024 / 025, Loss: 0.119188

* Train accuracy / confusion: 93.62% / [[342, 8, 3], [13, 252, 5], [7, 15, 155]], 
* Val   accuracy / confusion: 50.96% / [[28, 13, 5], [6, 17, 12], [6, 9, 8]]

------------------------------ Epoch 322 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.111838
- Iter 024 / 025, Loss: 0.225323

* Train accuracy / confusion: 95.12% / [[349, 8, 2], [19, 240, 5], [1, 4, 172]], 
* Val   accuracy / confusion: 59.62% / [[33, 5, 8], [7, 21, 7], [8, 7, 8]]

------------------------------ Epoch 323 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.202991
- Iter 024 / 025, Loss: 0.188438

* Train accuracy / confusion: 94.88% / [[343, 11, 1], [10, 255, 5], [3, 11, 161]], 
* Val   accuracy / confusion: 52.88% / [[28, 8, 10], [11, 18, 6], [8, 6, 9]]

------------------------------ Epoch 324 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.096175
- Iter 024 / 025, Loss: 0.158382

* Train accuracy / confusion: 96.38% / [[352, 3, 1], [7, 255, 6], [8, 4, 164]], 
* Val   accuracy / confusion: 60.58% / [[30, 8, 8], [9, 20, 6], [6, 4, 13]]

------------------------------ Epoch 325 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.126330
- Iter 024 / 025, Loss: 0.400828

* Train accuracy / confusion: 95.00% / [[343, 11, 3], [12, 247, 6], [4, 4, 170]], 
* Val   accuracy / confusion: 50.96% / [[31, 6, 9], [11, 14, 10], [8, 7, 8]]

------------------------------ Epoch 326 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.268011
- Iter 024 / 025, Loss: 0.175550

* Train accuracy / confusion: 94.88% / [[342, 9, 1], [14, 253, 6], [4, 7, 164]], 
* Val   accuracy / confusion: 58.65% / [[31, 11, 4], [6, 21, 8], [8, 6, 9]]

------------------------------ Epoch 327 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.058393
- Iter 024 / 025, Loss: 0.130238

* Train accuracy / confusion: 96.88% / [[355, 4, 2], [7, 251, 5], [3, 4, 169]], 
* Val   accuracy / confusion: 56.73% / [[32, 6, 8], [8, 19, 8], [10, 5, 8]]

------------------------------ Epoch 328 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.089376
- Iter 024 / 025, Loss: 0.101837

* Train accuracy / confusion: 94.62% / [[341, 10, 3], [15, 247, 6], [3, 6, 169]], 
* Val   accuracy / confusion: 53.85% / [[25, 15, 6], [5, 23, 7], [8, 7, 8]]

------------------------------ Epoch 329 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.074658
- Iter 024 / 025, Loss: 0.189727

* Train accuracy / confusion: 95.25% / [[345, 8, 2], [11, 254, 3], [2, 12, 163]], 
* Val   accuracy / confusion: 55.77% / [[33, 3, 10], [7, 17, 11], [10, 5, 8]]

------------------------------ Epoch 330 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.118538
- Iter 024 / 025, Loss: 0.143940

* Train accuracy / confusion: 93.75% / [[343, 10, 2], [15, 247, 6], [3, 14, 160]], 
* Val   accuracy / confusion: 57.69% / [[30, 10, 6], [10, 20, 5], [9, 4, 10]]

------------------------------ Epoch 331 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.157145
- Iter 024 / 025, Loss: 0.243839

* Train accuracy / confusion: 95.00% / [[346, 11, 3], [13, 252, 5], [5, 3, 162]], 
* Val   accuracy / confusion: 56.73% / [[30, 11, 5], [7, 21, 7], [7, 8, 8]]

------------------------------ Epoch 332 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.185401
- Iter 024 / 025, Loss: 0.270150

* Train accuracy / confusion: 95.00% / [[344, 6, 3], [14, 250, 9], [3, 5, 166]], 
* Val   accuracy / confusion: 51.92% / [[32, 8, 6], [13, 14, 8], [8, 7, 8]]

------------------------------ Epoch 333 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.082551
- Iter 024 / 025, Loss: 0.206687

* Train accuracy / confusion: 95.50% / [[343, 9, 4], [8, 253, 7], [4, 4, 168]], 
* Val   accuracy / confusion: 55.77% / [[29, 11, 6], [4, 22, 9], [6, 10, 7]]

------------------------------ Epoch 334 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.121970
- Iter 024 / 025, Loss: 0.099827

* Train accuracy / confusion: 95.62% / [[348, 7, 4], [10, 251, 5], [4, 5, 166]], 
* Val   accuracy / confusion: 50.96% / [[30, 9, 7], [6, 18, 11], [8, 10, 5]]

------------------------------ Epoch 335 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.060857
- Iter 024 / 025, Loss: 0.134461

* Train accuracy / confusion: 96.00% / [[348, 7, 2], [6, 254, 6], [3, 8, 166]], 
* Val   accuracy / confusion: 57.69% / [[31, 6, 9], [8, 17, 10], [5, 6, 12]]

------------------------------ Epoch 336 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.114968
- Iter 024 / 025, Loss: 0.120069

* Train accuracy / confusion: 94.50% / [[345, 11, 1], [15, 251, 5], [6, 6, 160]], 
* Val   accuracy / confusion: 50.00% / [[28, 9, 9], [12, 14, 9], [7, 6, 10]]

------------------------------ Epoch 337 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.115627
- Iter 024 / 025, Loss: 0.144535

* Train accuracy / confusion: 95.25% / [[342, 14, 1], [9, 257, 5], [2, 7, 163]], 
* Val   accuracy / confusion: 64.42% / [[35, 6, 5], [4, 21, 10], [6, 6, 11]]

------------------------------ Epoch 338 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.170041
- Iter 024 / 025, Loss: 0.153140

* Train accuracy / confusion: 95.62% / [[347, 6, 3], [9, 256, 5], [0, 12, 162]], 
* Val   accuracy / confusion: 55.77% / [[28, 13, 5], [7, 21, 7], [8, 6, 9]]

------------------------------ Epoch 339 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.161766
- Iter 024 / 025, Loss: 0.139191

* Train accuracy / confusion: 97.00% / [[349, 5, 1], [6, 256, 6], [2, 4, 171]], 
* Val   accuracy / confusion: 52.88% / [[31, 10, 5], [6, 17, 12], [7, 9, 7]]

------------------------------ Epoch 340 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.053955
- Iter 024 / 025, Loss: 0.149239

* Train accuracy / confusion: 95.38% / [[347, 9, 4], [8, 258, 6], [5, 5, 158]], 
* Val   accuracy / confusion: 59.62% / [[30, 11, 5], [5, 23, 7], [8, 6, 9]]

------------------------------ Epoch 341 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.074289
- Iter 024 / 025, Loss: 0.058394

* Train accuracy / confusion: 93.75% / [[339, 6, 6], [13, 251, 5], [6, 14, 160]], 
* Val   accuracy / confusion: 57.69% / [[29, 9, 8], [8, 21, 6], [5, 8, 10]]

------------------------------ Epoch 342 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.085902
- Iter 024 / 025, Loss: 0.122542

* Train accuracy / confusion: 95.25% / [[349, 9, 0], [16, 244, 5], [3, 5, 169]], 
* Val   accuracy / confusion: 49.04% / [[29, 9, 8], [10, 13, 12], [7, 7, 9]]

------------------------------ Epoch 343 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.102244
- Iter 024 / 025, Loss: 0.121246

* Train accuracy / confusion: 95.50% / [[348, 6, 6], [15, 242, 6], [1, 2, 174]], 
* Val   accuracy / confusion: 54.81% / [[28, 12, 6], [7, 22, 6], [8, 8, 7]]

------------------------------ Epoch 344 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.057329
- Iter 024 / 025, Loss: 0.125622

* Train accuracy / confusion: 96.00% / [[348, 7, 2], [15, 251, 3], [1, 4, 169]], 
* Val   accuracy / confusion: 49.04% / [[30, 10, 6], [11, 15, 9], [6, 11, 6]]

------------------------------ Epoch 345 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.107146
- Iter 024 / 025, Loss: 0.242954

* Train accuracy / confusion: 95.62% / [[346, 8, 4], [11, 252, 6], [3, 3, 167]], 
* Val   accuracy / confusion: 52.88% / [[29, 7, 10], [6, 19, 10], [8, 8, 7]]

------------------------------ Epoch 346 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.111245
- Iter 024 / 025, Loss: 0.102141

* Train accuracy / confusion: 95.00% / [[342, 12, 3], [12, 251, 5], [4, 4, 167]], 
* Val   accuracy / confusion: 53.85% / [[30, 9, 7], [9, 19, 7], [7, 9, 7]]

------------------------------ Epoch 347 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.047012
- Iter 024 / 025, Loss: 0.092968

* Train accuracy / confusion: 95.88% / [[347, 9, 2], [9, 254, 4], [4, 5, 166]], 
* Val   accuracy / confusion: 57.69% / [[31, 9, 6], [7, 19, 9], [10, 3, 10]]

------------------------------ Epoch 348 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.084608
- Iter 024 / 025, Loss: 0.147789

* Train accuracy / confusion: 95.38% / [[345, 9, 2], [14, 250, 3], [3, 6, 168]], 
* Val   accuracy / confusion: 57.69% / [[30, 10, 6], [6, 21, 8], [5, 9, 9]]

------------------------------ Epoch 349 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.137435
- Iter 024 / 025, Loss: 0.149188

* Train accuracy / confusion: 96.25% / [[351, 5, 2], [12, 252, 6], [3, 2, 167]], 
* Val   accuracy / confusion: 53.85% / [[31, 10, 5], [8, 16, 11], [7, 7, 9]]

------------------------------ Epoch 350 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.115971
- Iter 024 / 025, Loss: 0.135343

* Train accuracy / confusion: 96.00% / [[353, 6, 1], [9, 251, 7], [6, 3, 164]], 
* Val   accuracy / confusion: 51.92% / [[30, 9, 7], [5, 16, 14], [9, 6, 8]]

------------------------------ Epoch 351 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.132164
- Iter 024 / 025, Loss: 0.052679

* Train accuracy / confusion: 96.75% / [[353, 1, 3], [11, 247, 7], [3, 1, 174]], 
* Val   accuracy / confusion: 57.69% / [[31, 7, 8], [4, 19, 12], [6, 7, 10]]

------------------------------ Epoch 352 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.109779
- Iter 024 / 025, Loss: 0.061004

* Train accuracy / confusion: 95.38% / [[346, 8, 3], [9, 254, 7], [5, 5, 163]], 
* Val   accuracy / confusion: 52.88% / [[33, 9, 4], [11, 15, 9], [10, 6, 7]]

------------------------------ Epoch 353 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.167425
- Iter 024 / 025, Loss: 0.103746

* Train accuracy / confusion: 94.75% / [[346, 7, 1], [12, 248, 9], [5, 8, 164]], 
* Val   accuracy / confusion: 54.81% / [[30, 9, 7], [9, 20, 6], [9, 7, 7]]

------------------------------ Epoch 354 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.083282
- Iter 024 / 025, Loss: 0.198554

* Train accuracy / confusion: 94.25% / [[350, 8, 1], [11, 241, 12], [9, 5, 163]], 
* Val   accuracy / confusion: 49.04% / [[28, 9, 9], [9, 16, 10], [7, 9, 7]]

------------------------------ Epoch 355 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.066200
- Iter 024 / 025, Loss: 0.239460

* Train accuracy / confusion: 95.50% / [[351, 4, 4], [13, 249, 6], [4, 5, 164]], 
* Val   accuracy / confusion: 50.96% / [[32, 7, 7], [7, 13, 15], [8, 7, 8]]

------------------------------ Epoch 356 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.114296
- Iter 024 / 025, Loss: 0.043558

* Train accuracy / confusion: 93.88% / [[341, 13, 1], [12, 249, 9], [4, 10, 161]], 
* Val   accuracy / confusion: 51.92% / [[30, 10, 6], [11, 12, 12], [7, 4, 12]]

------------------------------ Epoch 357 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.189213
- Iter 024 / 025, Loss: 0.038253

* Train accuracy / confusion: 95.00% / [[340, 12, 3], [16, 251, 4], [3, 2, 169]], 
* Val   accuracy / confusion: 53.85% / [[31, 11, 4], [9, 16, 10], [5, 9, 9]]

------------------------------ Epoch 358 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.068104
- Iter 024 / 025, Loss: 0.284399

* Train accuracy / confusion: 95.12% / [[348, 7, 1], [14, 248, 4], [6, 7, 165]], 
* Val   accuracy / confusion: 57.69% / [[25, 13, 8], [6, 23, 6], [7, 4, 12]]

------------------------------ Epoch 359 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.131917
- Iter 024 / 025, Loss: 0.117974

* Train accuracy / confusion: 96.12% / [[350, 3, 2], [12, 248, 10], [3, 1, 171]], 
* Val   accuracy / confusion: 49.04% / [[26, 14, 6], [9, 18, 8], [9, 7, 7]]

------------------------------ Epoch 360 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.052601
- Iter 024 / 025, Loss: 0.170685

* Train accuracy / confusion: 93.88% / [[337, 13, 3], [17, 247, 7], [4, 5, 167]], 
* Val   accuracy / confusion: 57.69% / [[32, 8, 6], [10, 19, 6], [10, 4, 9]]

------------------------------ Epoch 361 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.077862
- Iter 024 / 025, Loss: 0.097370

* Train accuracy / confusion: 95.62% / [[347, 6, 3], [10, 253, 3], [6, 7, 165]], 
* Val   accuracy / confusion: 55.77% / [[30, 11, 5], [9, 20, 6], [8, 7, 8]]

------------------------------ Epoch 362 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.166725
- Iter 024 / 025, Loss: 0.058187

* Train accuracy / confusion: 95.62% / [[347, 6, 0], [15, 250, 5], [5, 4, 168]], 
* Val   accuracy / confusion: 50.96% / [[30, 10, 6], [9, 16, 10], [7, 9, 7]]

------------------------------ Epoch 363 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.130123
- Iter 024 / 025, Loss: 0.095109

* Train accuracy / confusion: 95.38% / [[345, 7, 2], [9, 249, 9], [2, 8, 169]], 
* Val   accuracy / confusion: 52.88% / [[26, 13, 7], [6, 20, 9], [3, 11, 9]]

------------------------------ Epoch 364 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.380151
- Iter 024 / 025, Loss: 0.188802

* Train accuracy / confusion: 95.12% / [[342, 9, 3], [13, 254, 3], [4, 7, 165]], 
* Val   accuracy / confusion: 53.85% / [[30, 9, 7], [8, 16, 11], [7, 6, 10]]

------------------------------ Epoch 365 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.106559
- Iter 024 / 025, Loss: 0.101736

* Train accuracy / confusion: 95.12% / [[343, 13, 1], [12, 251, 8], [1, 4, 167]], 
* Val   accuracy / confusion: 51.92% / [[29, 10, 7], [7, 18, 10], [9, 7, 7]]

------------------------------ Epoch 366 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.127207
- Iter 024 / 025, Loss: 0.198978

* Train accuracy / confusion: 95.50% / [[344, 11, 2], [14, 247, 4], [2, 3, 173]], 
* Val   accuracy / confusion: 51.92% / [[26, 13, 7], [6, 20, 9], [8, 7, 8]]

------------------------------ Epoch 367 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.157759
- Iter 024 / 025, Loss: 0.066283

* Train accuracy / confusion: 94.50% / [[340, 15, 3], [7, 258, 6], [5, 8, 158]], 
* Val   accuracy / confusion: 56.73% / [[30, 12, 4], [8, 17, 10], [4, 7, 12]]

------------------------------ Epoch 368 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.177843
- Iter 024 / 025, Loss: 0.087080

* Train accuracy / confusion: 95.38% / [[345, 6, 0], [12, 251, 8], [4, 7, 167]], 
* Val   accuracy / confusion: 55.77% / [[30, 9, 7], [10, 17, 8], [7, 5, 11]]

------------------------------ Epoch 369 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.110210
- Iter 024 / 025, Loss: 0.197185

* Train accuracy / confusion: 95.38% / [[344, 9, 5], [10, 254, 4], [4, 5, 165]], 
* Val   accuracy / confusion: 52.88% / [[32, 6, 8], [10, 16, 9], [6, 10, 7]]

------------------------------ Epoch 370 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.072842
- Iter 024 / 025, Loss: 0.077917

* Train accuracy / confusion: 94.62% / [[340, 9, 7], [9, 251, 7], [5, 6, 166]], 
* Val   accuracy / confusion: 54.81% / [[29, 10, 7], [8, 18, 9], [8, 5, 10]]

------------------------------ Epoch 371 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.165201
- Iter 024 / 025, Loss: 0.050801

* Train accuracy / confusion: 95.75% / [[343, 7, 6], [11, 252, 5], [4, 1, 171]], 
* Val   accuracy / confusion: 59.62% / [[32, 8, 6], [6, 19, 10], [6, 6, 11]]

------------------------------ Epoch 372 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.095647
- Iter 024 / 025, Loss: 0.264234

* Train accuracy / confusion: 95.88% / [[345, 5, 3], [15, 252, 3], [1, 6, 170]], 
* Val   accuracy / confusion: 51.92% / [[27, 11, 8], [8, 18, 9], [7, 7, 9]]

------------------------------ Epoch 373 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.103976
- Iter 024 / 025, Loss: 0.129242

* Train accuracy / confusion: 96.12% / [[347, 4, 5], [9, 256, 6], [4, 3, 166]], 
* Val   accuracy / confusion: 58.65% / [[25, 15, 6], [6, 22, 7], [6, 3, 14]]

------------------------------ Epoch 374 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.074319
- Iter 024 / 025, Loss: 0.099085

* Train accuracy / confusion: 96.50% / [[350, 6, 1], [7, 253, 6], [3, 5, 169]], 
* Val   accuracy / confusion: 54.81% / [[26, 13, 7], [7, 23, 5], [8, 7, 8]]

------------------------------ Epoch 375 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.121722
- Iter 024 / 025, Loss: 0.062893

* Train accuracy / confusion: 95.75% / [[347, 11, 1], [10, 251, 3], [4, 5, 168]], 
* Val   accuracy / confusion: 56.73% / [[31, 10, 5], [5, 20, 10], [6, 9, 8]]

------------------------------ Epoch 376 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.078494
- Iter 024 / 025, Loss: 0.039506

* Train accuracy / confusion: 96.38% / [[342, 5, 5], [6, 259, 7], [1, 5, 170]], 
* Val   accuracy / confusion: 54.81% / [[28, 9, 9], [6, 19, 10], [6, 7, 10]]

------------------------------ Epoch 377 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.162994
- Iter 024 / 025, Loss: 0.077847

* Train accuracy / confusion: 96.25% / [[351, 2, 2], [11, 253, 6], [1, 8, 166]], 
* Val   accuracy / confusion: 57.69% / [[32, 6, 8], [5, 20, 10], [9, 6, 8]]

------------------------------ Epoch 378 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.058640
- Iter 024 / 025, Loss: 0.076773

* Train accuracy / confusion: 95.25% / [[344, 8, 2], [12, 253, 4], [4, 8, 165]], 
* Val   accuracy / confusion: 52.88% / [[30, 9, 7], [5, 13, 17], [6, 5, 12]]

------------------------------ Epoch 379 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.073974
- Iter 024 / 025, Loss: 0.178405

* Train accuracy / confusion: 95.50% / [[347, 4, 3], [7, 252, 10], [4, 8, 165]], 
* Val   accuracy / confusion: 53.85% / [[29, 12, 5], [7, 21, 7], [5, 12, 6]]

------------------------------ Epoch 380 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.094522
- Iter 024 / 025, Loss: 0.113311

* Train accuracy / confusion: 96.50% / [[345, 8, 2], [6, 262, 3], [2, 7, 165]], 
* Val   accuracy / confusion: 57.69% / [[32, 6, 8], [8, 19, 8], [9, 5, 9]]

------------------------------ Epoch 381 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.172466
- Iter 024 / 025, Loss: 0.111935

* Train accuracy / confusion: 96.38% / [[342, 4, 4], [11, 256, 6], [1, 3, 173]], 
* Val   accuracy / confusion: 53.85% / [[29, 9, 8], [6, 18, 11], [9, 5, 9]]

------------------------------ Epoch 382 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.227618
- Iter 024 / 025, Loss: 0.089322

* Train accuracy / confusion: 94.12% / [[344, 11, 2], [16, 244, 6], [2, 10, 165]], 
* Val   accuracy / confusion: 54.81% / [[30, 11, 5], [10, 19, 6], [7, 8, 8]]

------------------------------ Epoch 383 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.084217
- Iter 024 / 025, Loss: 0.127281

* Train accuracy / confusion: 95.50% / [[346, 6, 1], [10, 249, 8], [4, 7, 169]], 
* Val   accuracy / confusion: 50.00% / [[28, 12, 6], [9, 18, 8], [11, 6, 6]]

------------------------------ Epoch 384 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.035183
- Iter 024 / 025, Loss: 0.070172

* Train accuracy / confusion: 95.62% / [[345, 10, 2], [12, 252, 4], [3, 4, 168]], 
* Val   accuracy / confusion: 50.96% / [[28, 10, 8], [9, 16, 10], [8, 6, 9]]

------------------------------ Epoch 385 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.144273
- Iter 024 / 025, Loss: 0.070165

* Train accuracy / confusion: 95.00% / [[350, 7, 2], [14, 247, 7], [3, 7, 163]], 
* Val   accuracy / confusion: 55.77% / [[31, 7, 8], [4, 18, 13], [8, 6, 9]]

------------------------------ Epoch 386 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.118377
- Iter 024 / 025, Loss: 0.398747

* Train accuracy / confusion: 96.62% / [[348, 4, 2], [8, 258, 4], [4, 5, 167]], 
* Val   accuracy / confusion: 56.73% / [[33, 6, 7], [7, 17, 11], [9, 5, 9]]

------------------------------ Epoch 387 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.085693
- Iter 024 / 025, Loss: 0.169555

* Train accuracy / confusion: 96.50% / [[349, 5, 2], [10, 255, 4], [2, 5, 168]], 
* Val   accuracy / confusion: 51.92% / [[30, 7, 9], [8, 18, 9], [9, 8, 6]]

------------------------------ Epoch 388 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.302511
- Iter 024 / 025, Loss: 0.137151

* Train accuracy / confusion: 96.00% / [[345, 7, 3], [11, 254, 4], [2, 5, 169]], 
* Val   accuracy / confusion: 53.85% / [[28, 12, 6], [6, 18, 11], [6, 7, 10]]

------------------------------ Epoch 389 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.146440
- Iter 024 / 025, Loss: 0.129617

* Train accuracy / confusion: 95.25% / [[343, 9, 2], [14, 252, 4], [5, 4, 167]], 
* Val   accuracy / confusion: 53.85% / [[27, 11, 8], [4, 21, 10], [4, 11, 8]]

------------------------------ Epoch 390 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.208799
- Iter 024 / 025, Loss: 0.132662

* Train accuracy / confusion: 94.75% / [[346, 7, 3], [7, 256, 7], [6, 12, 156]], 
* Val   accuracy / confusion: 52.88% / [[28, 11, 7], [8, 18, 9], [7, 7, 9]]

------------------------------ Epoch 391 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.067174
- Iter 024 / 025, Loss: 0.092678

* Train accuracy / confusion: 96.25% / [[350, 6, 1], [11, 252, 5], [3, 4, 168]], 
* Val   accuracy / confusion: 52.88% / [[32, 9, 5], [8, 14, 13], [6, 8, 9]]

------------------------------ Epoch 392 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.184368
- Iter 024 / 025, Loss: 0.136566

* Train accuracy / confusion: 95.75% / [[344, 8, 1], [11, 250, 7], [6, 1, 172]], 
* Val   accuracy / confusion: 50.00% / [[30, 9, 7], [7, 14, 14], [7, 8, 8]]

------------------------------ Epoch 393 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.109558
- Iter 024 / 025, Loss: 0.110908

* Train accuracy / confusion: 96.75% / [[351, 7, 0], [7, 250, 5], [2, 5, 173]], 
* Val   accuracy / confusion: 51.92% / [[26, 12, 8], [8, 18, 9], [7, 6, 10]]

------------------------------ Epoch 394 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.081464
- Iter 024 / 025, Loss: 0.049167

* Train accuracy / confusion: 95.75% / [[348, 5, 2], [9, 248, 12], [2, 4, 170]], 
* Val   accuracy / confusion: 50.00% / [[32, 8, 6], [8, 15, 12], [8, 10, 5]]

------------------------------ Epoch 395 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.059465
- Iter 024 / 025, Loss: 0.062041

* Train accuracy / confusion: 96.75% / [[354, 7, 0], [7, 257, 2], [3, 7, 163]], 
* Val   accuracy / confusion: 55.77% / [[31, 8, 7], [7, 21, 7], [7, 10, 6]]

------------------------------ Epoch 396 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.122119
- Iter 024 / 025, Loss: 0.143237

* Train accuracy / confusion: 96.38% / [[347, 5, 1], [12, 252, 4], [3, 4, 172]], 
* Val   accuracy / confusion: 51.92% / [[30, 10, 6], [6, 17, 12], [10, 6, 7]]

------------------------------ Epoch 397 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.095157
- Iter 024 / 025, Loss: 0.048567

* Train accuracy / confusion: 96.62% / [[348, 6, 1], [10, 251, 6], [2, 2, 174]], 
* Val   accuracy / confusion: 55.77% / [[30, 9, 7], [8, 16, 11], [5, 6, 12]]

------------------------------ Epoch 398 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.113079
- Iter 024 / 025, Loss: 0.074293

* Train accuracy / confusion: 94.62% / [[343, 8, 4], [14, 246, 7], [3, 7, 168]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [9, 17, 9], [9, 6, 8]]

------------------------------ Epoch 399 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.067166
- Iter 024 / 025, Loss: 0.087538

* Train accuracy / confusion: 95.38% / [[347, 7, 2], [9, 253, 5], [3, 11, 163]], 
* Val   accuracy / confusion: 50.00% / [[25, 16, 5], [7, 19, 9], [8, 7, 8]]

------------------------------ Epoch 400 / 500, Learning rate: 4.31e-05 ------------------------------
- Iter 012 / 025, Loss: 0.125027
- Iter 024 / 025, Loss: 0.099352

* Train accuracy / confusion: 95.50% / [[342, 7, 0], [15, 253, 4], [2, 8, 169]], 
* Val   accuracy / confusion: 50.00% / [[25, 14, 7], [8, 18, 9], [10, 4, 9]]

------------------------------ Epoch 401 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.073694
- Iter 024 / 025, Loss: 0.079904

* Train accuracy / confusion: 97.00% / [[346, 4, 1], [9, 257, 5], [0, 5, 173]], 
* Val   accuracy / confusion: 56.73% / [[31, 8, 7], [10, 19, 6], [7, 7, 9]]

------------------------------ Epoch 402 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.178382
- Iter 024 / 025, Loss: 0.085014

* Train accuracy / confusion: 95.75% / [[353, 3, 2], [10, 248, 9], [1, 9, 165]], 
* Val   accuracy / confusion: 56.73% / [[28, 12, 6], [6, 22, 7], [6, 8, 9]]

------------------------------ Epoch 403 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.101926
- Iter 024 / 025, Loss: 0.081046

* Train accuracy / confusion: 95.25% / [[347, 10, 2], [12, 248, 3], [2, 9, 167]], 
* Val   accuracy / confusion: 53.85% / [[28, 10, 8], [6, 22, 7], [7, 10, 6]]

------------------------------ Epoch 404 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.193774
- Iter 024 / 025, Loss: 0.103928

* Train accuracy / confusion: 95.88% / [[347, 11, 0], [9, 250, 7], [1, 5, 170]], 
* Val   accuracy / confusion: 50.00% / [[28, 12, 6], [12, 13, 10], [7, 5, 11]]

------------------------------ Epoch 405 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.045805
- Iter 024 / 025, Loss: 0.103879

* Train accuracy / confusion: 96.62% / [[351, 7, 1], [12, 253, 3], [2, 2, 169]], 
* Val   accuracy / confusion: 54.81% / [[29, 11, 6], [6, 19, 10], [6, 8, 9]]

------------------------------ Epoch 406 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.152230
- Iter 024 / 025, Loss: 0.126633

* Train accuracy / confusion: 95.25% / [[344, 9, 1], [12, 249, 7], [5, 4, 169]], 
* Val   accuracy / confusion: 53.85% / [[28, 10, 8], [8, 16, 11], [4, 7, 12]]

------------------------------ Epoch 407 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.109714
- Iter 024 / 025, Loss: 0.181848

* Train accuracy / confusion: 96.62% / [[348, 6, 1], [9, 258, 3], [4, 4, 167]], 
* Val   accuracy / confusion: 53.85% / [[28, 10, 8], [6, 18, 11], [9, 4, 10]]

------------------------------ Epoch 408 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.105019
- Iter 024 / 025, Loss: 0.164671

* Train accuracy / confusion: 97.25% / [[350, 5, 2], [9, 257, 2], [0, 4, 171]], 
* Val   accuracy / confusion: 54.81% / [[29, 11, 6], [7, 19, 9], [5, 9, 9]]

------------------------------ Epoch 409 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.032551
- Iter 024 / 025, Loss: 0.097030

* Train accuracy / confusion: 96.88% / [[351, 8, 1], [6, 256, 4], [5, 1, 168]], 
* Val   accuracy / confusion: 58.65% / [[31, 8, 7], [8, 22, 5], [10, 5, 8]]

------------------------------ Epoch 410 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.149815
- Iter 024 / 025, Loss: 0.089060

* Train accuracy / confusion: 96.00% / [[348, 7, 0], [12, 251, 5], [4, 4, 169]], 
* Val   accuracy / confusion: 58.65% / [[33, 5, 8], [8, 18, 9], [8, 5, 10]]

------------------------------ Epoch 411 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.078346
- Iter 024 / 025, Loss: 0.111257

* Train accuracy / confusion: 95.50% / [[350, 8, 1], [8, 252, 6], [6, 7, 162]], 
* Val   accuracy / confusion: 51.92% / [[28, 14, 4], [6, 19, 10], [9, 7, 7]]

------------------------------ Epoch 412 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.032351
- Iter 024 / 025, Loss: 0.042200

* Train accuracy / confusion: 95.75% / [[352, 7, 2], [7, 252, 8], [3, 7, 162]], 
* Val   accuracy / confusion: 52.88% / [[28, 10, 8], [10, 17, 8], [7, 6, 10]]

------------------------------ Epoch 413 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.082589
- Iter 024 / 025, Loss: 0.112619

* Train accuracy / confusion: 95.38% / [[343, 10, 2], [10, 252, 7], [5, 3, 168]], 
* Val   accuracy / confusion: 53.85% / [[28, 11, 7], [6, 19, 10], [6, 8, 9]]

------------------------------ Epoch 414 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.111824
- Iter 024 / 025, Loss: 0.129942

* Train accuracy / confusion: 96.62% / [[353, 6, 0], [6, 254, 5], [2, 8, 166]], 
* Val   accuracy / confusion: 54.81% / [[29, 10, 7], [8, 20, 7], [6, 9, 8]]

------------------------------ Epoch 415 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.055678
- Iter 024 / 025, Loss: 0.113813

* Train accuracy / confusion: 95.62% / [[346, 11, 2], [5, 252, 6], [5, 6, 167]], 
* Val   accuracy / confusion: 52.88% / [[28, 11, 7], [9, 19, 7], [9, 6, 8]]

------------------------------ Epoch 416 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.215226
- Iter 024 / 025, Loss: 0.175909

* Train accuracy / confusion: 95.62% / [[346, 8, 2], [14, 254, 3], [3, 5, 165]], 
* Val   accuracy / confusion: 56.73% / [[29, 10, 7], [11, 19, 5], [9, 3, 11]]

------------------------------ Epoch 417 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.078589
- Iter 024 / 025, Loss: 0.282701

* Train accuracy / confusion: 95.62% / [[346, 7, 1], [16, 248, 5], [2, 4, 171]], 
* Val   accuracy / confusion: 48.08% / [[25, 12, 9], [7, 18, 10], [10, 6, 7]]

------------------------------ Epoch 418 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.151950
- Iter 024 / 025, Loss: 0.053974

* Train accuracy / confusion: 96.75% / [[351, 3, 2], [12, 253, 3], [2, 4, 170]], 
* Val   accuracy / confusion: 50.96% / [[26, 15, 5], [7, 19, 9], [7, 8, 8]]

------------------------------ Epoch 419 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.248404
- Iter 024 / 025, Loss: 0.051672

* Train accuracy / confusion: 96.62% / [[349, 1, 3], [8, 257, 6], [3, 6, 167]], 
* Val   accuracy / confusion: 58.65% / [[31, 10, 5], [7, 20, 8], [8, 5, 10]]

------------------------------ Epoch 420 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.098483
- Iter 024 / 025, Loss: 0.072732

* Train accuracy / confusion: 96.50% / [[354, 4, 1], [7, 253, 8], [5, 3, 165]], 
* Val   accuracy / confusion: 50.00% / [[30, 8, 8], [9, 18, 8], [8, 11, 4]]

------------------------------ Epoch 421 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.081759
- Iter 024 / 025, Loss: 0.054004

* Train accuracy / confusion: 95.00% / [[344, 6, 4], [13, 249, 7], [5, 5, 167]], 
* Val   accuracy / confusion: 52.88% / [[28, 11, 7], [9, 20, 6], [9, 7, 7]]

------------------------------ Epoch 422 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.132661
- Iter 024 / 025, Loss: 0.228732

* Train accuracy / confusion: 96.62% / [[350, 3, 1], [8, 258, 5], [7, 3, 165]], 
* Val   accuracy / confusion: 54.81% / [[29, 8, 9], [8, 20, 7], [11, 4, 8]]

------------------------------ Epoch 423 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.117748
- Iter 024 / 025, Loss: 0.212897

* Train accuracy / confusion: 96.00% / [[350, 5, 2], [13, 251, 5], [2, 5, 167]], 
* Val   accuracy / confusion: 48.08% / [[26, 11, 9], [9, 18, 8], [8, 9, 6]]

------------------------------ Epoch 424 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.191398
- Iter 024 / 025, Loss: 0.053683

* Train accuracy / confusion: 97.00% / [[350, 5, 0], [10, 257, 1], [4, 4, 169]], 
* Val   accuracy / confusion: 50.00% / [[25, 12, 9], [8, 20, 7], [6, 10, 7]]

------------------------------ Epoch 425 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.121984
- Iter 024 / 025, Loss: 0.111294

* Train accuracy / confusion: 96.88% / [[346, 6, 2], [9, 259, 4], [2, 2, 170]], 
* Val   accuracy / confusion: 47.12% / [[26, 13, 7], [8, 14, 13], [6, 8, 9]]

------------------------------ Epoch 426 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.076836
- Iter 024 / 025, Loss: 0.190447

* Train accuracy / confusion: 96.88% / [[352, 3, 2], [9, 252, 5], [0, 6, 171]], 
* Val   accuracy / confusion: 54.81% / [[28, 12, 6], [7, 21, 7], [3, 12, 8]]

------------------------------ Epoch 427 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.351087
- Iter 024 / 025, Loss: 0.099661

* Train accuracy / confusion: 95.50% / [[344, 9, 3], [7, 251, 9], [5, 3, 169]], 
* Val   accuracy / confusion: 50.96% / [[27, 15, 4], [8, 17, 10], [9, 5, 9]]

------------------------------ Epoch 428 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.091797
- Iter 024 / 025, Loss: 0.111000

* Train accuracy / confusion: 96.12% / [[350, 9, 3], [10, 248, 4], [0, 5, 171]], 
* Val   accuracy / confusion: 53.85% / [[27, 13, 6], [10, 18, 7], [8, 4, 11]]

------------------------------ Epoch 429 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.161917
- Iter 024 / 025, Loss: 0.089620

* Train accuracy / confusion: 96.62% / [[355, 4, 1], [12, 246, 6], [1, 3, 172]], 
* Val   accuracy / confusion: 53.85% / [[30, 10, 6], [10, 16, 9], [6, 7, 10]]

------------------------------ Epoch 430 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.155186
- Iter 024 / 025, Loss: 0.086721

* Train accuracy / confusion: 95.88% / [[344, 9, 0], [9, 258, 2], [8, 5, 165]], 
* Val   accuracy / confusion: 55.77% / [[30, 9, 7], [8, 20, 7], [5, 10, 8]]

------------------------------ Epoch 431 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.143666
- Iter 024 / 025, Loss: 0.118537

* Train accuracy / confusion: 96.62% / [[346, 7, 4], [6, 261, 5], [2, 3, 166]], 
* Val   accuracy / confusion: 54.81% / [[27, 9, 10], [10, 19, 6], [5, 7, 11]]

------------------------------ Epoch 432 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.035752
- Iter 024 / 025, Loss: 0.109011

* Train accuracy / confusion: 95.88% / [[351, 6, 1], [12, 248, 4], [5, 5, 168]], 
* Val   accuracy / confusion: 52.88% / [[28, 13, 5], [6, 19, 10], [3, 12, 8]]

------------------------------ Epoch 433 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.057360
- Iter 024 / 025, Loss: 0.109753

* Train accuracy / confusion: 96.25% / [[349, 5, 1], [7, 258, 7], [5, 5, 163]], 
* Val   accuracy / confusion: 53.85% / [[31, 8, 7], [9, 16, 10], [9, 5, 9]]

------------------------------ Epoch 434 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.112236
- Iter 024 / 025, Loss: 0.087915

* Train accuracy / confusion: 96.62% / [[351, 6, 2], [6, 254, 7], [2, 4, 168]], 
* Val   accuracy / confusion: 49.04% / [[25, 13, 8], [8, 18, 9], [8, 7, 8]]

------------------------------ Epoch 435 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.133621
- Iter 024 / 025, Loss: 0.081721

* Train accuracy / confusion: 96.50% / [[347, 5, 3], [11, 255, 3], [2, 4, 170]], 
* Val   accuracy / confusion: 56.73% / [[32, 10, 4], [6, 19, 10], [10, 5, 8]]

------------------------------ Epoch 436 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.087955
- Iter 024 / 025, Loss: 0.127361

* Train accuracy / confusion: 96.75% / [[351, 5, 1], [8, 256, 4], [3, 5, 167]], 
* Val   accuracy / confusion: 56.73% / [[29, 10, 7], [9, 19, 7], [7, 5, 11]]

------------------------------ Epoch 437 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.176332
- Iter 024 / 025, Loss: 0.067846

* Train accuracy / confusion: 97.00% / [[346, 8, 2], [6, 257, 3], [1, 4, 173]], 
* Val   accuracy / confusion: 48.08% / [[28, 12, 6], [10, 14, 11], [7, 8, 8]]

------------------------------ Epoch 438 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.067039
- Iter 024 / 025, Loss: 0.114942

* Train accuracy / confusion: 97.00% / [[353, 5, 1], [7, 254, 5], [0, 6, 169]], 
* Val   accuracy / confusion: 50.96% / [[30, 10, 6], [9, 17, 9], [6, 11, 6]]

------------------------------ Epoch 439 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.160691
- Iter 024 / 025, Loss: 0.052096

* Train accuracy / confusion: 95.38% / [[344, 8, 2], [12, 252, 7], [2, 6, 167]], 
* Val   accuracy / confusion: 48.08% / [[25, 13, 8], [7, 17, 11], [6, 9, 8]]

------------------------------ Epoch 440 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.103585
- Iter 024 / 025, Loss: 0.262624

* Train accuracy / confusion: 96.88% / [[346, 7, 2], [7, 257, 7], [1, 1, 172]], 
* Val   accuracy / confusion: 51.92% / [[29, 8, 9], [10, 15, 10], [5, 8, 10]]

------------------------------ Epoch 441 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.064708
- Iter 024 / 025, Loss: 0.067039

* Train accuracy / confusion: 97.38% / [[350, 5, 1], [4, 260, 5], [4, 2, 169]], 
* Val   accuracy / confusion: 50.96% / [[28, 12, 6], [9, 17, 9], [7, 8, 8]]

------------------------------ Epoch 442 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.079896
- Iter 024 / 025, Loss: 0.141338

* Train accuracy / confusion: 95.62% / [[350, 8, 0], [17, 247, 4], [2, 4, 168]], 
* Val   accuracy / confusion: 59.62% / [[34, 7, 5], [4, 20, 11], [8, 7, 8]]

------------------------------ Epoch 443 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.148900
- Iter 024 / 025, Loss: 0.072205

* Train accuracy / confusion: 96.75% / [[345, 6, 3], [6, 259, 4], [1, 6, 170]], 
* Val   accuracy / confusion: 53.85% / [[31, 9, 6], [9, 17, 9], [6, 9, 8]]

------------------------------ Epoch 444 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.106783
- Iter 024 / 025, Loss: 0.046799

* Train accuracy / confusion: 96.62% / [[352, 5, 2], [10, 255, 3], [1, 6, 166]], 
* Val   accuracy / confusion: 56.73% / [[27, 13, 6], [6, 21, 8], [6, 6, 11]]

------------------------------ Epoch 445 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.068280
- Iter 024 / 025, Loss: 0.119808

* Train accuracy / confusion: 96.88% / [[354, 7, 2], [9, 256, 3], [1, 3, 165]], 
* Val   accuracy / confusion: 50.00% / [[26, 10, 10], [7, 18, 10], [8, 7, 8]]

------------------------------ Epoch 446 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.314783
- Iter 024 / 025, Loss: 0.089792

* Train accuracy / confusion: 95.50% / [[347, 7, 3], [12, 256, 4], [3, 7, 161]], 
* Val   accuracy / confusion: 51.92% / [[30, 9, 7], [9, 14, 12], [8, 5, 10]]

------------------------------ Epoch 447 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.032766
- Iter 024 / 025, Loss: 0.093899

* Train accuracy / confusion: 96.62% / [[347, 4, 0], [10, 256, 7], [4, 2, 170]], 
* Val   accuracy / confusion: 53.85% / [[29, 11, 6], [7, 16, 12], [6, 6, 11]]

------------------------------ Epoch 448 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.058991
- Iter 024 / 025, Loss: 0.338905

* Train accuracy / confusion: 95.75% / [[348, 5, 5], [10, 256, 3], [4, 7, 162]], 
* Val   accuracy / confusion: 51.92% / [[27, 11, 8], [4, 18, 13], [7, 7, 9]]

------------------------------ Epoch 449 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.343943
- Iter 024 / 025, Loss: 0.111069

* Train accuracy / confusion: 96.25% / [[347, 12, 0], [11, 256, 0], [3, 4, 167]], 
* Val   accuracy / confusion: 52.88% / [[28, 12, 6], [11, 18, 6], [8, 6, 9]]

------------------------------ Epoch 450 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.116977
- Iter 024 / 025, Loss: 0.100719

* Train accuracy / confusion: 96.88% / [[355, 3, 2], [8, 257, 4], [4, 4, 163]], 
* Val   accuracy / confusion: 49.04% / [[27, 11, 8], [10, 16, 9], [8, 7, 8]]

------------------------------ Epoch 451 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.084641
- Iter 024 / 025, Loss: 0.120614

* Train accuracy / confusion: 96.50% / [[349, 5, 1], [8, 254, 7], [0, 7, 169]], 
* Val   accuracy / confusion: 53.85% / [[29, 11, 6], [8, 21, 6], [8, 9, 6]]

------------------------------ Epoch 452 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.138829
- Iter 024 / 025, Loss: 0.147442

* Train accuracy / confusion: 97.12% / [[354, 3, 2], [6, 251, 6], [3, 3, 172]], 
* Val   accuracy / confusion: 47.12% / [[27, 11, 8], [10, 12, 13], [10, 3, 10]]

------------------------------ Epoch 453 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.117362
- Iter 024 / 025, Loss: 0.178897

* Train accuracy / confusion: 96.75% / [[349, 6, 2], [9, 254, 3], [4, 2, 171]], 
* Val   accuracy / confusion: 52.88% / [[27, 10, 9], [8, 18, 9], [7, 6, 10]]

------------------------------ Epoch 454 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.127288
- Iter 024 / 025, Loss: 0.208537

* Train accuracy / confusion: 97.50% / [[349, 5, 1], [7, 258, 5], [1, 1, 173]], 
* Val   accuracy / confusion: 61.54% / [[34, 7, 5], [6, 18, 11], [4, 7, 12]]

------------------------------ Epoch 455 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.075965
- Iter 024 / 025, Loss: 0.081482

* Train accuracy / confusion: 95.50% / [[345, 10, 2], [6, 252, 7], [2, 9, 167]], 
* Val   accuracy / confusion: 48.08% / [[29, 7, 10], [11, 15, 9], [12, 5, 6]]

------------------------------ Epoch 456 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.038763
- Iter 024 / 025, Loss: 0.168481

* Train accuracy / confusion: 96.50% / [[352, 5, 2], [9, 252, 3], [2, 7, 168]], 
* Val   accuracy / confusion: 53.85% / [[27, 13, 6], [7, 18, 10], [5, 7, 11]]

------------------------------ Epoch 457 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.128954
- Iter 024 / 025, Loss: 0.192844

* Train accuracy / confusion: 95.50% / [[350, 3, 4], [11, 250, 6], [4, 8, 164]], 
* Val   accuracy / confusion: 54.81% / [[28, 10, 8], [10, 19, 6], [7, 6, 10]]

------------------------------ Epoch 458 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.116034
- Iter 024 / 025, Loss: 0.202915

* Train accuracy / confusion: 95.38% / [[347, 9, 3], [6, 252, 7], [4, 8, 164]], 
* Val   accuracy / confusion: 50.00% / [[24, 15, 7], [7, 17, 11], [8, 4, 11]]

------------------------------ Epoch 459 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.050096
- Iter 024 / 025, Loss: 0.143539

* Train accuracy / confusion: 97.25% / [[353, 2, 3], [7, 260, 2], [2, 6, 165]], 
* Val   accuracy / confusion: 53.85% / [[27, 11, 8], [5, 22, 8], [9, 7, 7]]

------------------------------ Epoch 460 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.107897
- Iter 024 / 025, Loss: 0.148885

* Train accuracy / confusion: 96.62% / [[346, 7, 3], [8, 258, 0], [3, 6, 169]], 
* Val   accuracy / confusion: 51.92% / [[31, 9, 6], [13, 13, 9], [5, 8, 10]]

------------------------------ Epoch 461 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.032967
- Iter 024 / 025, Loss: 0.136169

* Train accuracy / confusion: 95.50% / [[349, 8, 1], [9, 255, 4], [6, 8, 160]], 
* Val   accuracy / confusion: 61.54% / [[32, 8, 6], [7, 23, 5], [8, 6, 9]]

------------------------------ Epoch 462 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.210906
- Iter 024 / 025, Loss: 0.052900

* Train accuracy / confusion: 95.50% / [[341, 9, 1], [12, 251, 8], [1, 5, 172]], 
* Val   accuracy / confusion: 50.00% / [[28, 11, 7], [12, 14, 9], [7, 6, 10]]

------------------------------ Epoch 463 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.086030
- Iter 024 / 025, Loss: 0.060770

* Train accuracy / confusion: 96.12% / [[342, 9, 1], [8, 256, 5], [2, 6, 171]], 
* Val   accuracy / confusion: 54.81% / [[31, 8, 7], [7, 17, 11], [8, 6, 9]]

------------------------------ Epoch 464 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.152775
- Iter 024 / 025, Loss: 0.049038

* Train accuracy / confusion: 95.88% / [[340, 8, 4], [8, 254, 4], [5, 4, 173]], 
* Val   accuracy / confusion: 51.92% / [[30, 8, 8], [7, 16, 12], [8, 7, 8]]

------------------------------ Epoch 465 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.137176
- Iter 024 / 025, Loss: 0.326120

* Train accuracy / confusion: 95.12% / [[347, 9, 3], [16, 246, 4], [2, 5, 168]], 
* Val   accuracy / confusion: 48.08% / [[29, 6, 11], [11, 14, 10], [11, 5, 7]]

------------------------------ Epoch 466 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.068016
- Iter 024 / 025, Loss: 0.080773

* Train accuracy / confusion: 96.88% / [[351, 5, 1], [10, 257, 2], [1, 6, 167]], 
* Val   accuracy / confusion: 54.81% / [[27, 11, 8], [5, 20, 10], [6, 7, 10]]

------------------------------ Epoch 467 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.047694
- Iter 024 / 025, Loss: 0.068976

* Train accuracy / confusion: 96.12% / [[341, 12, 2], [7, 257, 4], [2, 4, 171]], 
* Val   accuracy / confusion: 55.77% / [[32, 8, 6], [8, 16, 11], [7, 6, 10]]

------------------------------ Epoch 468 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.168117
- Iter 024 / 025, Loss: 0.173229

* Train accuracy / confusion: 96.25% / [[354, 3, 4], [9, 252, 3], [5, 6, 164]], 
* Val   accuracy / confusion: 59.62% / [[33, 6, 7], [6, 19, 10], [6, 7, 10]]

------------------------------ Epoch 469 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.085982
- Iter 024 / 025, Loss: 0.353771

* Train accuracy / confusion: 96.62% / [[352, 4, 3], [8, 253, 6], [1, 5, 168]], 
* Val   accuracy / confusion: 52.88% / [[32, 8, 6], [10, 14, 11], [8, 6, 9]]

------------------------------ Epoch 470 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.071757
- Iter 024 / 025, Loss: 0.188379

* Train accuracy / confusion: 96.88% / [[346, 4, 3], [7, 260, 6], [0, 5, 169]], 
* Val   accuracy / confusion: 57.69% / [[32, 10, 4], [7, 20, 8], [10, 5, 8]]

------------------------------ Epoch 471 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.141407
- Iter 024 / 025, Loss: 0.241776

* Train accuracy / confusion: 96.00% / [[352, 5, 2], [9, 246, 8], [2, 6, 170]], 
* Val   accuracy / confusion: 53.85% / [[30, 8, 8], [9, 19, 7], [8, 8, 7]]

------------------------------ Epoch 472 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.139051
- Iter 024 / 025, Loss: 0.274025

* Train accuracy / confusion: 96.38% / [[343, 7, 4], [8, 258, 3], [6, 1, 170]], 
* Val   accuracy / confusion: 51.92% / [[26, 12, 8], [9, 18, 8], [6, 7, 10]]

------------------------------ Epoch 473 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.050914
- Iter 024 / 025, Loss: 0.039267

* Train accuracy / confusion: 96.00% / [[350, 5, 3], [10, 253, 6], [3, 5, 165]], 
* Val   accuracy / confusion: 53.85% / [[32, 5, 9], [12, 18, 5], [12, 5, 6]]

------------------------------ Epoch 474 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.152743
- Iter 024 / 025, Loss: 0.129112

* Train accuracy / confusion: 97.25% / [[348, 4, 2], [2, 264, 5], [4, 5, 166]], 
* Val   accuracy / confusion: 50.96% / [[26, 14, 6], [9, 19, 7], [7, 8, 8]]

------------------------------ Epoch 475 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.096435
- Iter 024 / 025, Loss: 0.112521

* Train accuracy / confusion: 96.25% / [[343, 12, 1], [6, 255, 7], [2, 2, 172]], 
* Val   accuracy / confusion: 52.88% / [[28, 9, 9], [5, 19, 11], [9, 6, 8]]

------------------------------ Epoch 476 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.042361
- Iter 024 / 025, Loss: 0.127975

* Train accuracy / confusion: 96.62% / [[345, 6, 2], [11, 257, 5], [1, 2, 171]], 
* Val   accuracy / confusion: 50.00% / [[26, 11, 9], [7, 19, 9], [7, 9, 7]]

------------------------------ Epoch 477 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.057192
- Iter 024 / 025, Loss: 0.071864

* Train accuracy / confusion: 96.25% / [[345, 9, 1], [9, 252, 5], [3, 3, 173]], 
* Val   accuracy / confusion: 51.92% / [[28, 11, 7], [8, 18, 9], [6, 9, 8]]

------------------------------ Epoch 478 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.073333
- Iter 024 / 025, Loss: 0.111690

* Train accuracy / confusion: 96.62% / [[344, 6, 3], [11, 256, 3], [2, 2, 173]], 
* Val   accuracy / confusion: 50.96% / [[29, 9, 8], [9, 19, 7], [10, 8, 5]]

------------------------------ Epoch 479 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.079332
- Iter 024 / 025, Loss: 0.248195

* Train accuracy / confusion: 96.25% / [[349, 6, 1], [7, 254, 5], [3, 8, 167]], 
* Val   accuracy / confusion: 52.88% / [[30, 11, 5], [6, 18, 11], [11, 5, 7]]

------------------------------ Epoch 480 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.063242
- Iter 024 / 025, Loss: 0.101763

* Train accuracy / confusion: 97.38% / [[352, 4, 2], [7, 254, 3], [2, 3, 173]], 
* Val   accuracy / confusion: 54.81% / [[28, 9, 9], [10, 19, 6], [7, 6, 10]]

------------------------------ Epoch 481 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.080918
- Iter 024 / 025, Loss: 0.102372

* Train accuracy / confusion: 95.50% / [[350, 7, 1], [13, 253, 3], [6, 6, 161]], 
* Val   accuracy / confusion: 57.69% / [[31, 11, 4], [5, 20, 10], [7, 7, 9]]

------------------------------ Epoch 482 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.082442
- Iter 024 / 025, Loss: 0.175353

* Train accuracy / confusion: 95.62% / [[351, 10, 0], [10, 245, 7], [4, 4, 169]], 
* Val   accuracy / confusion: 55.77% / [[28, 10, 8], [9, 20, 6], [6, 7, 10]]

------------------------------ Epoch 483 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.160220
- Iter 024 / 025, Loss: 0.108212

* Train accuracy / confusion: 96.75% / [[356, 3, 3], [12, 252, 0], [4, 4, 166]], 
* Val   accuracy / confusion: 54.81% / [[30, 8, 8], [3, 18, 14], [5, 9, 9]]

------------------------------ Epoch 484 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.054876
- Iter 024 / 025, Loss: 0.209407

* Train accuracy / confusion: 96.62% / [[357, 3, 1], [6, 253, 4], [5, 8, 163]], 
* Val   accuracy / confusion: 50.00% / [[26, 13, 7], [8, 16, 11], [8, 5, 10]]

------------------------------ Epoch 485 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.183597
- Iter 024 / 025, Loss: 0.304546

* Train accuracy / confusion: 96.62% / [[351, 5, 0], [7, 258, 6], [3, 6, 164]], 
* Val   accuracy / confusion: 51.92% / [[28, 10, 8], [7, 17, 11], [9, 5, 9]]

------------------------------ Epoch 486 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.048444
- Iter 024 / 025, Loss: 0.220994

* Train accuracy / confusion: 95.88% / [[340, 10, 4], [7, 259, 3], [4, 5, 168]], 
* Val   accuracy / confusion: 54.81% / [[29, 10, 7], [9, 17, 9], [7, 5, 11]]

------------------------------ Epoch 487 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.102796
- Iter 024 / 025, Loss: 0.068726

* Train accuracy / confusion: 95.88% / [[342, 8, 4], [8, 256, 5], [1, 7, 169]], 
* Val   accuracy / confusion: 53.85% / [[31, 8, 7], [11, 14, 10], [6, 6, 11]]

------------------------------ Epoch 488 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.054121
- Iter 024 / 025, Loss: 0.050982

* Train accuracy / confusion: 96.88% / [[347, 5, 1], [8, 258, 4], [4, 3, 170]], 
* Val   accuracy / confusion: 54.81% / [[26, 15, 5], [4, 22, 9], [8, 6, 9]]

------------------------------ Epoch 489 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.442185
- Iter 024 / 025, Loss: 0.140167

* Train accuracy / confusion: 96.50% / [[344, 8, 2], [3, 258, 8], [2, 5, 170]], 
* Val   accuracy / confusion: 54.81% / [[29, 9, 8], [8, 15, 12], [5, 5, 13]]

------------------------------ Epoch 490 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.194559
- Iter 024 / 025, Loss: 0.257415

* Train accuracy / confusion: 95.50% / [[346, 8, 6], [10, 249, 7], [2, 3, 169]], 
* Val   accuracy / confusion: 55.77% / [[31, 9, 6], [9, 17, 9], [8, 5, 10]]

------------------------------ Epoch 491 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.155985
- Iter 024 / 025, Loss: 0.148403

* Train accuracy / confusion: 96.50% / [[347, 5, 3], [8, 258, 4], [4, 4, 167]], 
* Val   accuracy / confusion: 51.92% / [[29, 11, 6], [7, 19, 9], [9, 8, 6]]

------------------------------ Epoch 492 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.150475
- Iter 024 / 025, Loss: 0.249365

* Train accuracy / confusion: 97.00% / [[352, 4, 3], [8, 252, 5], [2, 2, 172]], 
* Val   accuracy / confusion: 54.81% / [[28, 8, 10], [7, 18, 10], [5, 7, 11]]

------------------------------ Epoch 493 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.132459
- Iter 024 / 025, Loss: 0.037293

* Train accuracy / confusion: 95.50% / [[353, 4, 2], [13, 252, 7], [5, 5, 159]], 
* Val   accuracy / confusion: 60.58% / [[33, 4, 9], [3, 24, 8], [8, 9, 6]]

------------------------------ Epoch 494 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.086901
- Iter 024 / 025, Loss: 0.176712

* Train accuracy / confusion: 96.38% / [[349, 2, 3], [10, 254, 5], [6, 3, 168]], 
* Val   accuracy / confusion: 56.73% / [[32, 8, 6], [9, 17, 9], [7, 6, 10]]

------------------------------ Epoch 495 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.091247
- Iter 024 / 025, Loss: 0.057687

* Train accuracy / confusion: 97.12% / [[349, 6, 1], [6, 259, 3], [2, 5, 169]], 
* Val   accuracy / confusion: 52.88% / [[29, 8, 9], [7, 16, 12], [8, 5, 10]]

------------------------------ Epoch 496 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.099665
- Iter 024 / 025, Loss: 0.286608

* Train accuracy / confusion: 96.50% / [[352, 5, 0], [10, 251, 6], [3, 4, 169]], 
* Val   accuracy / confusion: 50.00% / [[29, 10, 7], [9, 18, 8], [10, 8, 5]]

------------------------------ Epoch 497 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.145507
- Iter 024 / 025, Loss: 0.091063

* Train accuracy / confusion: 96.50% / [[346, 5, 3], [9, 256, 6], [0, 5, 170]], 
* Val   accuracy / confusion: 54.81% / [[30, 12, 4], [10, 20, 5], [5, 11, 7]]

------------------------------ Epoch 498 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.038297
- Iter 024 / 025, Loss: 0.247700

* Train accuracy / confusion: 95.88% / [[341, 8, 4], [10, 249, 10], [1, 0, 177]], 
* Val   accuracy / confusion: 51.92% / [[28, 11, 7], [9, 17, 9], [6, 8, 9]]

------------------------------ Epoch 499 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.136800
- Iter 024 / 025, Loss: 0.050619

* Train accuracy / confusion: 97.12% / [[352, 3, 2], [6, 254, 7], [0, 5, 171]], 
* Val   accuracy / confusion: 53.85% / [[25, 14, 7], [9, 19, 7], [6, 5, 12]]

------------------------------ Epoch 500 / 500, Learning rate: 4.31e-06 ------------------------------
- Iter 012 / 025, Loss: 0.073499
- Iter 024 / 025, Loss: 0.115009

* Train accuracy / confusion: 96.75% / [[348, 9, 0], [9, 256, 1], [4, 3, 170]], 
* Val   accuracy / confusion: 51.92% / [[30, 9, 7], [6, 16, 13], [6, 9, 8]]

**************************************** Training Ends ****************************************
- Test accuracy: 52.56%

- Confusion matrix:
 [[898 340 142]
 [252 418 350]
 [ 89 293 308]]

- Debug table:
{   '00941': {'GT': 0, 'Acc': ' 30.00%', 'Pred': [9, 20, 1]},
    '00408': {'GT': 0, 'Acc': ' 16.67%', 'Pred': [5, 25, 0]},
    '00871': {'GT': 0, 'Acc': ' 83.33%', 'Pred': [25, 5, 0]},
    '00852': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 3, 0]},
    '00816': {'GT': 0, 'Acc': ' 93.33%', 'Pred': [28, 2, 0]},
    '00672': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00414': {'GT': 2, 'Acc': ' 16.67%', 'Pred': [0, 25, 5]},
    '00767': {'GT': 1, 'Acc': ' 33.33%', 'Pred': [8, 10, 12]},
    '01148': {'GT': 1, 'Acc': ' 43.33%', 'Pred': [11, 13, 6]},
    '00743': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 0, 3]},
    '00305': {'GT': 1, 'Acc': ' 33.33%', 'Pred': [5, 10, 15]},
    '01305': {'GT': 1, 'Acc': ' 66.67%', 'Pred': [7, 20, 3]},
    '00455': {'GT': 1, 'Acc': ' 40.00%', 'Pred': [1, 12, 17]},
    '01140': {'GT': 0, 'Acc': ' 96.67%', 'Pred': [29, 0, 1]},
    '00649': {'GT': 2, 'Acc': ' 26.67%', 'Pred': [7, 15, 8]},
    '01068': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00176': {'GT': 0, 'Acc': ' 30.00%', 'Pred': [9, 16, 5]},
    '00861': {'GT': 2, 'Acc': ' 86.67%', 'Pred': [4, 0, 26]},
    '00302': {'GT': 1, 'Acc': '  0.00%', 'Pred': [0, 0, 30]},
    '00756': {'GT': 1, 'Acc': ' 36.67%', 'Pred': [15, 11, 4]},
    '00706': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01218': {'GT': 0, 'Acc': ' 36.67%', 'Pred': [11, 10, 9]},
    '01008': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00914': {'GT': 0, 'Acc': ' 43.33%', 'Pred': [13, 16, 1]},
    '00680': {'GT': 1, 'Acc': ' 73.33%', 'Pred': [8, 22, 0]},
    '00736': {'GT': 2, 'Acc': ' 23.33%', 'Pred': [6, 17, 7]},
    '00417': {'GT': 2, 'Acc': ' 73.33%', 'Pred': [0, 8, 22]},
    '01184': {'GT': 2, 'Acc': ' 73.33%', 'Pred': [0, 8, 22]},
    '00949': {'GT': 1, 'Acc': ' 36.67%', 'Pred': [6, 11, 13]},
    '01268': {'GT': 1, 'Acc': ' 53.33%', 'Pred': [5, 16, 9]},
    '00441': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00731': {'GT': 0, 'Acc': ' 60.00%', 'Pred': [18, 6, 6]},
    '00806': {'GT': 2, 'Acc': ' 23.33%', 'Pred': [4, 19, 7]},
    '00192': {'GT': 0, 'Acc': ' 70.00%', 'Pred': [21, 9, 0]},
    '01075': {'GT': 2, 'Acc': '  3.33%', 'Pred': [11, 18, 1]},
    '00058': {'GT': 0, 'Acc': ' 40.00%', 'Pred': [12, 3, 15]},
    '01126': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 3, 0]},
    '00924': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01070': {'GT': 0, 'Acc': ' 10.00%', 'Pred': [3, 27, 0]},
    '00901': {'GT': 0, 'Acc': ' 23.33%', 'Pred': [7, 19, 4]},
    '01079': {'GT': 1, 'Acc': '  0.00%', 'Pred': [26, 0, 4]},
    '00614': {'GT': 0, 'Acc': ' 93.33%', 'Pred': [28, 1, 1]},
    '00062': {'GT': 1, 'Acc': ' 70.00%', 'Pred': [0, 21, 9]},
    '00405': {'GT': 2, 'Acc': ' 63.33%', 'Pred': [1, 10, 19]},
    '01307': {'GT': 0, 'Acc': ' 46.67%', 'Pred': [14, 16, 0]},
    '00918': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 15, 15]},
    '00299': {'GT': 0, 'Acc': ' 23.33%', 'Pred': [7, 13, 10]},
    '01046': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01106': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00822': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '01037': {'GT': 1, 'Acc': ' 56.67%', 'Pred': [1, 17, 12]},
    '00599': {'GT': 1, 'Acc': ' 46.67%', 'Pred': [0, 14, 16]},
    '00517': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00832': {'GT': 2, 'Acc': ' 16.67%', 'Pred': [13, 12, 5]},
    '00498': {'GT': 1, 'Acc': ' 20.00%', 'Pred': [21, 6, 3]},
    '00280': {'GT': 2, 'Acc': ' 10.00%', 'Pred': [0, 27, 3]},
    '00623': {'GT': 2, 'Acc': ' 73.33%', 'Pred': [2, 6, 22]},
    '01038': {'GT': 1, 'Acc': ' 43.33%', 'Pred': [0, 13, 17]},
    '00534': {'GT': 0, 'Acc': ' 70.00%', 'Pred': [21, 9, 0]},
    '01199': {'GT': 2, 'Acc': ' 63.33%', 'Pred': [0, 11, 19]},
    '00886': {'GT': 0, 'Acc': ' 16.67%', 'Pred': [5, 22, 3]},
    '00698': {'GT': 1, 'Acc': ' 63.33%', 'Pred': [3, 19, 8]},
    '00509': {'GT': 0, 'Acc': ' 50.00%', 'Pred': [15, 1, 14]},
    '00389': {'GT': 1, 'Acc': ' 63.33%', 'Pred': [9, 19, 2]},
    '01208': {'GT': 2, 'Acc': '  0.00%', 'Pred': [1, 29, 0]},
    '00004': {'GT': 2, 'Acc': ' 83.33%', 'Pred': [0, 5, 25]},
    '01039': {'GT': 1, 'Acc': ' 80.00%', 'Pred': [0, 24, 6]},
    '01086': {'GT': 2, 'Acc': ' 40.00%', 'Pred': [9, 9, 12]},
    '00174': {'GT': 2, 'Acc': ' 96.67%', 'Pred': [1, 0, 29]},
    '00226': {'GT': 2, 'Acc': ' 90.00%', 'Pred': [0, 3, 27]},
    '00794': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00289': {'GT': 2, 'Acc': '  3.33%', 'Pred': [1, 28, 1]},
    '01027': {'GT': 0, 'Acc': ' 86.67%', 'Pred': [26, 4, 0]},
    '00855': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 3, 0]},
    '00385': {'GT': 0, 'Acc': '  0.00%', 'Pred': [0, 13, 17]},
    '00022': {'GT': 1, 'Acc': '  3.33%', 'Pred': [14, 1, 15]},
    '00220': {'GT': 1, 'Acc': ' 16.67%', 'Pred': [22, 5, 3]},
    '00635': {'GT': 2, 'Acc': '  3.33%', 'Pred': [27, 2, 1]},
    '00122': {'GT': 0, 'Acc': ' 76.67%', 'Pred': [23, 0, 7]},
    '00505': {'GT': 0, 'Acc': ' 60.00%', 'Pred': [18, 12, 0]},
    '01014': {'GT': 1, 'Acc': ' 20.00%', 'Pred': [24, 6, 0]},
    '00407': {'GT': 1, 'Acc': ' 10.00%', 'Pred': [24, 3, 3]},
    '01300': {'GT': 2, 'Acc': '  0.00%', 'Pred': [0, 30, 0]},
    '00301': {'GT': 1, 'Acc': ' 13.33%', 'Pred': [0, 4, 26]},
    '01349': {'GT': 1, 'Acc': '100.00%', 'Pred': [0, 30, 0]},
    '00005': {'GT': 2, 'Acc': ' 73.33%', 'Pred': [2, 6, 22]},
    '01169': {'GT': 0, 'Acc': ' 90.00%', 'Pred': [27, 1, 2]},
    '00124': {'GT': 0, 'Acc': '100.00%', 'Pred': [30, 0, 0]},
    '00243': {'GT': 1, 'Acc': ' 10.00%', 'Pred': [20, 3, 7]},
    '01240': {'GT': 1, 'Acc': ' 40.00%', 'Pred': [0, 12, 18]},
    '00589': {'GT': 0, 'Acc': ' 13.33%', 'Pred': [4, 16, 10]},
    '01172': {'GT': 1, 'Acc': '  0.00%', 'Pred': [0, 0, 30]},
    '01156': {'GT': 1, 'Acc': ' 96.67%', 'Pred': [1, 29, 0]},
    '00329': {'GT': 0, 'Acc': ' 10.00%', 'Pred': [3, 25, 2]},
    '01333': {'GT': 0, 'Acc': ' 56.67%', 'Pred': [17, 13, 0]},
    '01364': {'GT': 1, 'Acc': ' 50.00%', 'Pred': [0, 15, 15]},
    '00415': {'GT': 1, 'Acc': ' 83.33%', 'Pred': [3, 25, 2]},
    '00603': {'GT': 2, 'Acc': ' 83.33%', 'Pred': [0, 5, 25]},
    '00096': {'GT': 0, 'Acc': ' 46.67%', 'Pred': [14, 0, 16]},
    '01250': {'GT': 1, 'Acc': ' 40.00%', 'Pred': [14, 12, 4]},
    '00125': {'GT': 0, 'Acc': ' 60.00%', 'Pred': [18, 12, 0]},
    '00447': {'GT': 1, 'Acc': ' 43.33%', 'Pred': [1, 13, 16]},
    '00820': {'GT': 1, 'Acc': '  6.67%', 'Pred': [3, 2, 25]}}